{ "data": { "a": { "slug": "hy4", "name": "Tencent Hy4", "vendor": "Tencent (Hy team)", "vendorUrl": "https://aistudio.tencent.com", "kind": "model", "category": "inference", "summary": "Tencent Hy4 is Tencent's flagship language model, released on 4 November 2026 with image understanding and computer use, served on Tencent Cloud TokenHub as hy4 with a 1M-token context, at ¥6 input and ¥18 output per million tokens.", "url": "https://www.anchorterminal.com/tools/hy4", "markdownUrl": "https://www.anchorterminal.com/tools/hy4.md", "slimMarkdownUrl": "https://www.anchorterminal.com/tools/hy4.min.md", "jsonUrl": "https://www.anchorterminal.com/api/v1/tools/hy4.json", "dns": "hy4.grade.anchorterminal.com", "license": "Hosted model under the Tencent Cloud TokenHub terms. Tencent says weights and a system card will follow; no licence is stated yet", "transports": [ "http" ], "packages": [], "auth": "api-key", "authNotes": "A TokenHub bearer key (or `x-api-key` on the Anthropic-style route) from a Tencent Cloud account with real-name verification, per the tencent-tokenhub listing. Keys can be limited by model, service, IP range and token quota.", "pricing": "usage", "pricingNotes": "¥6 input, ¥0.3 cached input and ¥18 output per million tokens on TokenHub, billed at 50% until 31 December 2026 (Beijing time), per the price page updated 9 November 2026. Tencent's release gives about US$0.90, US$0.045 and US$2.70. No free allowance named for hy4 on the price page.", "priceSummary": "Pay per use", "where": "local", "x402": { "level": "no", "evidence": "No x402, MPP or L402 in Tencent's release or TokenHub's price page, 11 November 2026.", "endpoints": [] }, "toolCount": null, "popularity": { "githubStars": null, "npmWeekly": null, "pypiWeekly": null, "asOf": "2026-11-11" }, "docsUrl": "https://cloud.tencent.com/document/product/1823/130051", "llmsTxt": "https://cloud.tencent.com/document/product/1823/llms.txt", "capabilities": [ "inference.llm" ], "tags": [ "official", "hosted", "model", "multimodal", "reasoning", "long-context", "tool-calling", "usage-priced", "openai-compatible", "anthropic-compatible", "chinese", "new" ], "lastRelease": "2026-11-04", "graded": true, "anchor": { "graded": true, "score": 64, "grade": "B", "agentReady": false, "rank": 595, "ranked": true, "rankOf": 1524, "categoryRank": 25, "methodology": "0.8", "run": "2026-10-01", "scores": { "ergonomics": 86, "maintenance": 73, "payments": 45, "reliability": 46, "schema": 73, "security": 71, "transparency": 52 }, "pending": [ "performance", "tasks" ], "assessment": { "confidence": "medium", "date": "2026-11-11" }, "negative": 0, "verdict": "Tencent's flagship model is on TokenHub as hy4, with a 1M-token context, image understanding, function calling and caching, at ¥6 input and ¥18 output per million tokens, half price until the end of 2026. The weights it is called open source for aren't published yet, and every benchmark is Tencent's own.", "bestFor": "Long-context agent work, coding and office or finance tasks on a 1M-token context with images, for teams already on Tencent Cloud or happy to verify an account there.", "strengths": [ "1M-token context with 960K input and 64K output, per TokenHub's model list", "Image understanding, function calling, structured output, thinking and caching are marked for hy4", "Prices published: ¥6 input, ¥0.3 cached, ¥18 output per million tokens, billed at 50% until 31 December 2026", "OpenAI- and Anthropic-compatible endpoints on TokenHub, so existing SDKs work with a changed base URL", "Tencent says it's the same price as Hy4 preview despite the upgrade" ], "weaknesses": [ "Weights, licence and system card are promised but not published; tencent/Hy4 didn't answer on Hugging Face on 11 November", "Computer use, which Tencent's release names, isn't marked in TokenHub's model list", "TokenHub needs real-name verification, and the international site answers with a bot check", "No public status page for TokenHub", "Every benchmark in the release is Tencent's own or a partner's, several of them internal" ], "agentNotes": [ "Use model id hy4 on TokenHub's OpenAI- or Anthropic-compatible endpoints; hy4-preview is a separate, older model", "Keep input under 960K tokens and expect at most 64K output", "Reuse prompt prefixes: cached input is ¥0.3 against ¥6", "Check the price page for the end of the 50% discount on 31 December 2026 before budgeting", "Read rate limits from the console or DescribeModelQuota; the public docs don't give them for chat models" ], "metrics": { "kind": "remote", "measured": false }, "reviewCount": 0, "avgRating": 0, "history": [ { "basis": "public evidence", "confidence": "medium", "grade": "B", "methodology": "0.8", "pending": [ "performance", "tasks" ], "run": "2026-10-01", "runLabel": "October 2026 research run", "score": 64 } ], "editorialScores": { "ergonomics": 86, "maintenance": 73, "payments": 45, "reliability": 46, "schema": 73, "security": 71, "transparency": 55 }, "provenanceScore": 49, "checked": { "categories": { "ergonomics": "2026-11-11", "maintenance": "2026-11-11", "payments": "2026-11-11", "reliability": "2026-11-11", "schema": "2026-11-11", "security": "2026-11-11", "transparency": "2026-11-11" }, "last": "2026-11-11", "lastMoved": "2026-11-13", "oldest": "2026-11-11", "stale": false, "staleAfterDays": 90 } }, "connect": { "http": "curl https://tokenhub.tencentmaas.com/v1/chat/completions \\\n -H \"Authorization: Bearer $TOKENHUB_API_KEY\" -H \"content-type: application/json\" \\\n -d '{\"model\":\"hy4\",\"messages\":[{\"role\":\"user\",\"content\":\"hello\"}]}'" }, "letme": { "capability": "https://letme.dev/inference.llm", "tool": "https://letme.dev/hy4" }, "sameCompany": [ "hy-memory", "tencentdb-agent-memory", "hunyuanocr", "octop", "hy-memory-30b-a3b", "hy3", "wevisone-vl-2b", "wevisone-vl-4b" ], "sameModel": [ "tencent-tokenhub" ], "area": "models", "unitPrices": [ { "item": "Hy4 input", "unit": "1m-tokens", "usd": 0.9, "note": "¥6 list, about US$0.90 per Tencent; billed at 50% until 31 December 2026" }, { "item": "Hy4 cached input", "unit": "1m-tokens", "usd": 0.045, "note": "¥0.3 list" }, { "item": "Hy4 output", "unit": "1m-tokens", "usd": 2.7, "note": "¥18 list" } ], "provenance": { "legalEntity": "Tencent Cloud Computing (Beijing) Co., Ltd. (腾讯云计算（北京）有限责任公司)", "domain": "tencentmaas.com", "domainRegistered": "2026-03-13", "endpointOnVendorDomain": false, "terms": "https://cloud.tencent.com/document/product/301/129852", "privacy": "https://privacy.qq.com/document/preview/91d5fd3c98ce4468ae462ac59233bba2", "statusPage": "", "changelog": "https://cloud.tencent.com/document/product/1823/130675", "securityTxt": "none", "checked": "2026-11-11", "notes": [ "The TokenHub Terms of Service (effective 14 September 2026), the SLA (effective 23 April 2026) and the special note on third-party deployed models (effective 14 September 2026) name 腾讯云计算（北京）有限责任公司. The docs footer also names 深圳市腾讯计算机系统有限公司 as ICP licence holder.", "The TokenHub privacy policy is linked from the docs to privacy.qq.com, which renders only with JavaScript, so it was not read. privacy.qq.com has no robots.txt (404).", "No public status page was found. status.cloud.tencent.com did not answer, and the Tencent Cloud health dashboard documentation answered with a bot check.", "security.txt answered 404 on cloud.tencent.com, www.tencent.com and tokenhub.tencentmaas.com. Tencent's bug bounty policy at en.security.tencent.com lists Tencent Cloud in scope.", "RDAP at Verisign gives tencentmaas.com a registration date of 13 March 2026 with no registrant. The API answers at tokenhub.tencentmaas.com and tokenhub-intl.tencentmaas.com, named in Tencent Cloud's API usage page.", "cloud.tencent.com allows these docs in robots.txt. Its pages answered some requests with a JavaScript bot check, which was not passed. The pages were read when a later plain request returned them. www.tencentcloud.com answered robots.txt with a bot check, so the international site was not read." ], "score": 49 }, "pageJsonUrl": "https://www.anchorterminal.com/tools/hy4.json", "live": { "slug": "hy4", "securityTxt": { "url": "https://tencentmaas.com/.well-known/security.txt", "state": "unknown", "checkedAt": "2026-11-13T16:16:35.489580858Z" }, "llmsTxt": { "url": "https://cloud.tencent.com/document/product/1823/llms.txt", "ok": true, "status": 200, "checkedAt": "2026-11-13T15:54:34.939672287Z" }, "pages": [ { "url": "https://cloud.tencent.com/document/product/1823/130675", "kind": "changelog", "status": 200, "checkedAt": "2026-11-13T18:11:34.856577424Z", "changedAt": "0001-01-01T00:00:00Z", "fingerprint": "e3b0c44298fc" }, { "url": "https://privacy.qq.com/document/preview/91d5fd3c98ce4468ae462ac59233bba2", "kind": "privacy", "status": 304, "checkedAt": "2026-11-13T18:28:00.062507799Z", "changedAt": "0001-01-01T00:00:00Z", "fingerprint": "e3b0c44298fc" }, { "url": "https://cloud.tencent.com/document/product/301/129852", "kind": "terms", "status": 200, "checkedAt": "2026-11-13T18:11:36.909670614Z", "changedAt": "0001-01-01T00:00:00Z", "fingerprint": "e3b0c44298fc" } ], "updatedAt": "2026-11-13T18:28:00.062507799Z" } }, "answer": "Tencent Hy4 scores 64 (B) on agent readiness against Muse Spark Open's 56.5 (C), and leads in 5 of 7 scored categories. Muse Spark Open leads on payments \u0026 pricing.", "b": { "slug": "muse-spark-open", "name": "Muse Spark Open", "vendor": "Meta", "vendorUrl": "https://dev.meta.ai/models/muse-spark-open", "kind": "model", "category": "inference", "summary": "Muse Spark Open is Meta's open-weight multimodal model for coding and agentic work, with 816B total and 42B active parameters and a 1M-token context, released on 15 October 2026 under the Muse Model Licence for self-hosting.", "url": "https://www.anchorterminal.com/tools/muse-spark-open", "markdownUrl": "https://www.anchorterminal.com/tools/muse-spark-open.md", "slimMarkdownUrl": "https://www.anchorterminal.com/tools/muse-spark-open.min.md", "jsonUrl": "https://www.anchorterminal.com/api/v1/tools/muse-spark-open.json", "dns": "muse-spark-open.grade.anchorterminal.com", "repo": "https://huggingface.co/meta-models/Muse-Spark-Open-816B-A42B", "license": "Muse Model Licence (source-available weights, not an OSI licence), free of charge with an attribution notice, a separate agreement with Meta before Model as a Service by a licensee group above $20M gross revenue, and the Muse Spark Open Usage Policy", "transports": [ "http" ], "packages": [], "auth": "none", "authNotes": "Running the model needs no account or key. Getting Meta's weights does. Both Hugging Face repositories (BF16 and NVFP4) are gated with manual review and ask for full legal name, date of birth, country, affiliation and job title, and downloads then use a Hugging Face access token. Meta's Get the model page also gives a download from its own site after signing in and accepting the licence, one file at a time. The GGUF quantisations from ggml-org and Unsloth that Meta's docs point to are not gated. A local server has whatever authentication the owner adds, such as llama.cpp's `--api-key`.", "pricing": "free", "pricingNotes": "Free to download and run under the Muse Model Licence, with the owner's GPUs as the cost. A licensee whose group has more than $20M gross revenue in any 12 months must notify Meta and sign a separate agreement before offering it as Model as a Service. Hosted copies cost $1.25 input, $0.15 cached input and $4.25 output per million tokens on Fireworks, Baseten and Modal (checked 2026-10-15). On 4 November Hugging Face lists Baseten and Together as live inference providers, Baseten at $4.25 output. Meta Model API does not serve it.", "priceSummary": "from $1.25 / 1M in", "where": "local", "x402": { "level": "no", "evidence": "No x402, MPP or L402. Muse Spark Open is a model the owner runs, with no payment route from Meta, and Meta Model API does not serve it (checked 2026-10-15).", "endpoints": [] }, "toolCount": null, "popularity": { "githubStars": null, "npmWeekly": null, "pypiWeekly": null, "asOf": "2026-10-15" }, "docsUrl": "https://dev.meta.ai/docs/muse-spark-open", "llmsTxt": "https://dev.meta.ai/llms.txt", "capabilities": [ "inference.llm", "inference.open-weights", "inference.local" ], "tags": [ "model", "open-weights", "self-hosted", "local", "free", "gated", "multimodal", "openai-compatible", "llms-txt", "gpu", "new" ], "lastRelease": "2026-10-15", "graded": true, "anchor": { "graded": true, "score": 56.5, "grade": "C", "agentReady": false, "rank": 919, "ranked": true, "rankOf": 1524, "categoryRank": 33, "methodology": "0.8", "run": "2026-10-01", "scores": { "ergonomics": 69, "maintenance": 60, "payments": 75, "reliability": 42, "schema": 61, "security": 44, "transparency": 53 }, "pending": [ "performance", "tasks" ], "assessment": { "confidence": "medium", "date": "2026-11-04" }, "negative": 0, "verdict": "Muse Spark Open runs on the owner's hardware in seven runtimes, with OpenAI-compatible servers in vLLM and SGLang, standard tool calls, a 1M-token context and Meta's own serving docs. Serving the 537 GB NVFP4 checkpoint takes four B300 GPUs, Meta's download is gated by manual review, and the licence adds attribution and a revenue threshold for model resellers.", "bestFor": "A team with a four-GPU Blackwell node, or two to four Mac Studios, that wants a 1M-token multimodal coding and agent model on its own hardware with no prompts leaving it.", "strengths": [ "816B total and 42B active parameters with a 1,048,576-token context, taking text, image, video and audio input", "Day-one support in vLLM, SGLang, TensorRT-LLM, Transformers, llama.cpp, MLX and Unsloth, with commands for each in Meta's docs", "With the `muse` parsers, reasoning comes back in its own field and tool calls as a standard `tool_calls` array", "The licence grants use, fine-tuning, distillation and resale free of charge, with no user cap and no regional exclusion", "llms.txt and Markdown docs cover checkpoints, prompting, stop tokens, recommended settings and post-training" ], "weaknesses": [ "The NVFP4 checkpoint (about 537 GB) serves on four B300 GPUs and the BF16 one (about 1.65 TB) on eight", "Meta's Hugging Face repositories need manual approval with legal name and date of birth, which the Get the model page doesn't mention", "Products must display \"Built with Muse Spark Open\", and a company group with over $20M revenue needs Meta's agreement to offer it as Model as a Service", "The Usage Policy rules out military, nuclear, espionage, critical infrastructure and heavy machinery uses, and users under 18", "No changelog, deprecation policy or CI found, and none of the Hugging Face discussion replies counted on 4 November came from Meta" ], "agentNotes": [ "Start vLLM or SGLang with `--reasoning-parser muse --tool-call-parser muse` (vLLM also `--enable-auto-tool-choice`), or tool calls come back as raw text", "Set `reasoning_effort` to `low` or `medium` for short tasks. The default is `xhigh`, reasoning can't be turned off, and other values return an error", "Stop only on `\u003c|eot|\u003e` and end of text. Stopping on `\u003c|eom|\u003e` cuts off tool calls and final answers", "Stream long generations and leave `max_tokens` room for reasoning, or `content` can come back empty with `finish_reason` `length`", "Don't call Meta Model API for this model. Use a self-hosted server or a hosted provider's model string, such as `meta-models/muse-spark-open`" ], "metrics": { "kind": "local", "measured": false }, "reviewCount": 0, "avgRating": 0, "history": [ { "basis": "public evidence", "confidence": "medium", "grade": "C", "methodology": "0.8", "pending": [ "performance", "tasks" ], "run": "2026-10-01", "runLabel": "October 2026 research run", "score": 56.5 } ], "editorialScores": { "ergonomics": 69, "maintenance": 60, "payments": 75, "reliability": 42, "schema": 61, "security": 44, "transparency": 50 }, "provenanceScore": 55, "checked": { "categories": { "ergonomics": "2026-11-04", "maintenance": "2026-11-04", "payments": "2026-11-04", "reliability": "2026-11-04", "schema": "2026-11-04", "security": "2026-11-04", "transparency": "2026-11-04" }, "last": "2026-11-04", "lastMoved": "2026-11-13", "oldest": "2026-11-04", "stale": false, "staleAfterDays": 90 } }, "connect": { "install": "hf download meta-models/Muse-Spark-Open-816B-A42B-NVFP4 --local-dir ./muse-spark-open-nvfp4\n\nexport VLLM_USE_RUST_FRONTEND=1\nvllm serve meta-models/Muse-Spark-Open-816B-A42B-NVFP4 \\\n --tensor-parallel-size 4 \\\n --enable-expert-parallel \\\n --max-model-len 1048576 \\\n --reasoning-parser muse \\\n --tool-call-parser muse \\\n --enable-auto-tool-choice", "http": "curl http://localhost:8000/v1/chat/completions \\\n -H 'Content-Type: application/json' \\\n -d '{\n \"model\": \"meta-models/Muse-Spark-Open-816B-A42B-NVFP4\",\n \"messages\": [{\"role\": \"user\", \"content\": \"Refactor this module.\"}],\n \"reasoning_effort\": \"xhigh\"\n }'" }, "letme": { "capability": "https://letme.dev/inference.llm", "tool": "https://letme.dev/muse-spark-open" }, "sameCompany": [ "meta-model-api", "muse-ember", "llama-guard", "llamafirewall", "whatsapp-cloud-api", "muse-code", "meta-marketing-api", "muse-glimmer-30b" ], "family": { "lab": "Meta", "family": "Muse", "url": "https://www.anchorterminal.com/open-weights/#meta", "others": [ { "name": "Muse Glimmer 30B", "url": "https://huggingface.co/meta-models/Muse-Glimmer-30B", "params": "30B (dense, multimodal)", "listing": "https://www.anchorterminal.com/tools/muse-glimmer-30b" }, { "name": "Muse Glimmer 30B assistant (DFlash drafter)", "url": "https://huggingface.co/meta-models/Muse-Glimmer-30B-assistant", "params": "2.56B (Hub safetensors count; the card states no figure)" } ], "olderCount": 50, "unrestrictedCopies": 1 }, "area": "models", "provenance": { "legalEntity": "Meta Platforms, Inc.", "domain": "meta.ai", "domainRegistered": "2017-12-16", "endpointOnVendorDomain": null, "terms": "https://huggingface.co/meta-models/Muse-Spark-Open-816B-A42B/blob/main/LICENSE.md", "privacy": "", "statusPage": "", "changelog": "", "securityTxt": "none", "checked": "2026-11-04", "notes": [ "The licence names Meta Platforms, Inc. as copyright holder, and defines Meta as Meta Platforms Ireland Limited for licensees in the European region and Meta Platforms, Inc. elsewhere. The model card names the same two as model provider.", "The licence is the document that governs use of the weights, so it is recorded as the terms. LICENSE.md matches the text shown on the gated access form word for word. The Usage Policy is its Exhibit A. The separate USE_POLICY.md that Meta's docs link answered 401 without access and wasn't read.", "No privacy policy governs the model, because the owner runs it and no input reaches Meta. The access form says its details are handled under the Meta Privacy Policy on www.facebook.com, whose robots.txt asks readers like ours not to fetch it, so it wasn't read.", "The domain is meta.ai, where Meta's developer docs (dev.meta.ai) and research post (research.meta.ai) sit, the same as the meta-model-api and muse-code listings. RDAP gives a registration date of 16 December 2017.", "No hosted endpoint from Meta, so no status page. No changelog for the model was found, and the Hugging Face commit history needs access to read.", "dev.meta.ai/.well-known/security.txt and research.meta.ai/.well-known/security.txt returned 404 on 15 October 2026.", "WebFetch read the research post. It refused the Hugging Face model card as outside its allowed set, so the card, LICENSE.md, the repository and collection metadata, the discussions, the evaluation methodology PDF, Meta's docs pages and the providers' model pages were read with curl.", "Re-read on 4 November 2026 from the Hugging Face model pages, which show both Meta repositories gated with manual review and last changed on 15 October 2026. Meta's Get the model page now also lists a signed-in download from dev.meta.ai." ], "score": 55 }, "pageJsonUrl": "https://www.anchorterminal.com/tools/muse-spark-open.json", "live": { "slug": "muse-spark-open", "securityTxt": { "url": "https://meta.ai/.well-known/security.txt", "state": "none", "checkedAt": "2026-11-13T16:16:58.652426779Z" }, "llmsTxt": { "url": "https://dev.meta.ai/llms.txt", "ok": true, "status": 200, "checkedAt": "2026-11-13T15:55:19.454789941Z" }, "domain": { "domain": "meta.ai", "registered": "2017-12-16", "source": "https://rdap.identitydigital.services/rdap/domain/meta.ai", "checkedAt": "2026-11-08T14:51:01.706547979Z" }, "pages": [ { "url": "https://huggingface.co/meta-models/Muse-Spark-Open-816B-A42B/blob/main/LICENSE.md", "kind": "terms", "status": 200, "checkedAt": "2026-11-13T18:23:10.060930731Z", "changedAt": "2026-11-13T18:23:10.060930731Z", "fingerprint": "777eed6fee66" } ], "models": { "meta/muse-spark-open": { "openrouterId": "meta/muse-spark-open", "tools": true, "structuredOutputs": true, "jsonMode": true, "reasoning": true, "input": [ "audio", "file", "image", "text", "video" ], "promptCaching": true, "contextTokens": 1048576, "maxOutput": 943718, "checkedAt": "2026-11-13T23:49:32.552484543Z" } }, "updatedAt": "2026-11-13T23:49:32.552484543Z" } }, "edges": [ { "by": 30, "category": "payments", "leads": "muse-spark-open", "name": "Payments \u0026 pricing", "reasons": { "a": "Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. No machine payment protocol (0). The price page, updated 9 November, lists ¥6 input, ¥0.3 cached and ¥18 output per million tokens, with Hy4 billed at 50% until 31 December 2026 (20). TokenHub's campaign of 1M free tokens per language model, read on 1 November, isn't confirmed for hy4 on the price page, so half (10 of 20, a judgement call). A person signs up with real-name verification (0). 0 + 20 + 10 + 0 = 30. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 30 is 45.", "b": "Self-hosted rule. Nothing to buy from Meta, and the licence grants use free of charge. The Model as a Service clause asks for a separate agreement only from licensee groups above $20M gross revenue in 12 months (20). Free to run with no card (20). Meta's own weights need a manually reviewed Hugging Face form with legal name and date of birth, or, since the 4 November re-read of Get the model, a signed-in download from Meta's own page, so an agent can't get them alone. The ungated GGUF quantisations from ggml-org and Unsloth that Meta's docs point to download without approval (the Unsloth repository is ungated on 4 November), so part of the line (10 of 20). No payment protocol (0). Hosted copies at $1.25 input and $4.25 output per million tokens need each provider's own account and are not graded. 20 + 20 + 10 + 0 = 50. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 50 is 75." }, "scores": { "a": 45, "b": 75 } }, { "by": 27, "category": "security", "leads": "hy4", "name": "Security \u0026 auth", "reasons": { "a": "Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. Bearer keys limited by model, service, IP range and token quota, long-lived with no rotation (26 of 30). The terms say inputs aren't used to train Tencent Cloud's own models, and Hy4 is first party, so no third party's terms apply (14 of 20). Retention only as long as the service needs, no periods (7 of 15). Per-call audit logs and optional full logs to the customer's CLS (14 of 15). No security.txt; Tencent's TSRC bounty covers Tencent Cloud; no certification checked (10 of 20). 26 + 14 + 7 + 14 + 10 = 71.", "b": "Read as software the owner runs. No account or key to run it. Meta's weights need manual approval on Hugging Face and an access token, and server authentication is the owner's to add (12 of 30). An agent model that calls whatever tools the operator gives it. The card recommends human-in-the-loop confirmation for irreversible actions and Meta reports agentic-risk evaluations, but nothing in the model or template enforces an approval (8 of 20). The chat template neutralises special-token sequences and media placeholders in caller text and wraps tool results in `\u003ctool_output\u003e`, and the card advises protecting internal assistant messages and layering defences against adversarial input. No indirect prompt-injection results are published (10 of 15). No call log of its own. Visibility depends on the serving stack (5 of 15). Weights are safetensors that load through Transformers' own classes from 5.19, and Meta published cyber, chemical, biological and misalignment evaluations. dev.meta.ai/.well-known/security.txt answers 404 (re-read on 4 November), research.meta.ai per the 15 October check, and no security route is on the card or in the docs, only museoss@meta.com and the Hugging Face discussions (9 of 20). 12 + 8 + 10 + 5 + 9 = 44." }, "scores": { "a": 71, "b": 44 } }, { "by": 17, "category": "ergonomics", "leads": "hy4", "name": "Agent ergonomics", "reasons": { "a": "Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. Function calling is marked for hy4 in the model list (18 of 20). Structured output marked (13 of 15). Cache-hit price published, ¥0.3 against ¥6 input (14 of 15). 1M-token context with 960K input and 64K output (15). Batch at half price covered Hy3 and Hy4 preview on 1 November; whether it covers hy4 wasn't re-read (3 of 10). OpenAI and Anthropic SDKs with a changed base URL, plus Tencent Cloud's SDK and CLI for the control plane (8 of 10). Errors with code, messages, upstream status and request ID (15 of 15). 18 + 13 + 14 + 15 + 3 + 8 + 15 = 86.", "b": "Model reading of the checklist (tool use, structured output, caching, context, throughput, defaults and SDKs, errors). OpenAI-format tools and parallel calls come back as `tool_calls`, and the template supports namespaced tools and deferred loading through a `tool_search` tool. A forced `tool_choice` is not documented (17 of 20). JSON mode and JSON Schema come from each runtime's guided decoding, which Meta's docs show for TensorRT-LLM only (8 of 15). Prefix caching is the runtime's. Meta's SGLang command turns on hierarchical caching, and Dynamo routes each request to the worker holding its cached context (8 of 10). 1,048,576-token context, with sliding-window attention on 46 of 62 layers holding KV memory down (15 of 15). Draft models for speculative decoding in SGLang and vLLM, and no batch mode of its own (7 of 10). Any OpenAI SDK works against a local server, but reasoning can't be turned off, defaults to the most expensive level, and needs streaming, a `max_tokens` budget for reasoning and the right stop tokens (6 of 15). The template raises named errors for an unsupported reasoning effort, tool arguments passed as a string and a tool message with no name. Server errors are the runtime's (8 of 15). 17 + 8 + 8 + 15 + 7 + 6 + 8 = 69." }, "scores": { "a": 86, "b": 69 } }, { "by": 13, "category": "maintenance", "leads": "hy4", "name": "Maintenance \u0026 community", "reasons": { "a": "Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. Added on 4 November 2026 per the release notes; the model list and price page were updated on 9 November (30). Retirement notices of 21 to 33 days, no minimum written (7 of 12). Heavy retirement churn on the platform (3 of 8). Dated release notes and announcements, round-the-clock support (13 of 15). CLI replies not sampled (3 of 10). Control-plane SDKs and CLI, no inference SDK of its own (12 of 15). CLI runtime stated, CI not checked (5 of 10). 30 + 7 + 3 + 13 + 3 + 12 + 5 = 73. The model is a week old, so these lines are TokenHub's habits more than Hy4's record.", "b": "Read for an open-weight model released three weeks ago. Released on 15 October 2026 (30). One release, BF16 and NVFP4 on the same day, with no history of revisions, successors or notice periods, and Meta Model API has no deprecation policy either (6 of 20). None of the 29 replies counted by script on 4 November across the BF16 repository's discussions came from Meta. Feedback goes to the discussions or museoss@meta.com, and Meta's docs and developer blog carry dated posts (8 of 25). Day-one support in vLLM, SGLang, TensorRT-LLM, Transformers, llama.cpp, MLX and Unsloth with Meta's own recipes, though SGLang's is a dev image and TensorRT-LLM's a feature branch (12 of 15). Current dependencies (Transformers 5.19), but no CI visible to us (4 of 10). 30 + 6 + 8 + 12 + 4 = 60." }, "scores": { "a": 73, "b": 60 } }, { "by": 12, "category": "schema", "leads": "hy4", "name": "Schema \u0026 documentation", "reasons": { "a": "Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. No OpenAPI file for inference; field pages for the OpenAI and Anthropic formats as TokenHub implements them (12 of 25). llms.txt with Markdown copies (10). The model list, updated 9 November, gives hy4's context, input and output limits and marks thinking, structured output, function calling, caching and image understanding; computer use, which Tencent's release names, isn't marked (13 of 20). Typed field pages (10 of 15). Examples in five languages and an error page of 30 codes (14 of 15). Dated release notes, with hy4 added on 4 November (14 of 15). 12 + 10 + 13 + 10 + 14 + 14 = 73.", "b": "Read for a model the owner serves. No specification of its own. The contract is the Jinja chat template, which renders OpenAI-format tools with `additionalProperties: false` added and rejects bad values, plus the OpenAI-compatible servers in vLLM and SGLang with the `muse` parsers (10 of 25). dev.meta.ai/llms.txt lists the Muse Spark Open overview and seven pages under it, plus three cookbook pages, each with a Markdown copy (10). The card lists seven intended uses and six limitations, the docs say when to pick BF16 or NVFP4 and which reasoning strength suits which task, and the Usage Policy states what is out of scope (17 of 20). `reasoning_effort` is an enum of four values and anything else returns an error, and tools take JSON Schema. Structured output is documented only for TensorRT-LLM's `response_format`, not in the prompting guide or for vLLM and SGLang (11 of 15). curl and Python examples for every runtime and hosted provider, the rendered prompt for each case, and the template's own error messages, but no error reference for a served model (11 of 15). One release, with no changelog or version scheme beyond the licence's version line, Muse Spark Open (2 of 15). 10 + 10 + 17 + 11 + 11 + 2 = 61." }, "scores": { "a": 73, "b": 61 } } ], "facts": [ { "a": "Model API", "b": "Model API", "name": "Kind" }, { "a": "Tencent (Hy team)", "b": "Meta", "name": "Vendor" }, { "a": "no (local only)", "b": "no (local only)", "name": "Hosted endpoint" }, { "a": "HTTP", "b": "HTTP", "name": "Transports" }, { "a": "API key", "b": "None", "name": "Auth" }, { "a": "Pay per use", "b": "Free", "name": "Pricing" }, { "a": "no", "b": "no", "name": "x402" }, { "a": "Hosted model under the Tencent Cloud TokenHub terms. Tencent says weights and a system card will follow; no licence is stated yet", "b": "Muse Model Licence (source-available weights, not an OSI licence), free of charge with an attribution notice, a separate agreement with Meta before Model as a Service by a licensee group above $20M gross revenue, and the Muse Spark Open Usage Policy", "name": "Licence" }, { "a": "no", "b": "no", "name": "Read-only variant documented" }, { "a": "yes", "b": "yes", "name": "llms.txt" }, { "a": "2026-11-04", "b": "2026-10-15", "name": "Last release" }, { "a": "couldn't be read", "b": "2026-10-15", "name": "Terms last updated" }, { "a": "couldn't be read", "b": "no document linked", "name": "Privacy policy last updated" }, { "a": "couldn't be read", "b": "not found in the text", "name": "Customer content may train models" }, { "a": "couldn't be read", "b": "not found in the text", "name": "Terms restrict automated access" }, { "a": "couldn't be read", "b": "not found in the text", "name": "Terms restrict benchmarking" }, { "a": "couldn't be read", "b": "not found in the text", "name": "Terms or service can change without notice" }, { "a": "couldn't be read", "b": "not found in the text", "name": "Arbitration or class-action waiver" } ], "faq": [ { "answer": "Tencent Hy4 scores 64 (B) on agent readiness against Muse Spark Open's 56.5 (C), and leads in 5 of 7 scored categories. Muse Spark Open leads on payments \u0026 pricing.", "question": "Which is better for AI agents, Tencent Hy4 or Muse Spark Open?" }, { "answer": "Tencent Hy4 needs an API key. Muse Spark Open needs no key.", "question": "Do Tencent Hy4 and Muse Spark Open need an API key?" }, { "answer": "No hosted endpoint is listed for Tencent Hy4. No hosted endpoint is listed for Muse Spark Open.", "question": "Can an agent call Tencent Hy4 and Muse Spark Open without installing anything?" } ], "goodFor": [ { "aheadOn": [ "Schema \u0026 documentation, 73 against 61", "Agent ergonomics, 86 against 69", "Security \u0026 auth, 71 against 44", "Maintenance \u0026 community, 73 against 60" ], "also": null, "goodFor": "Long-context agent work, coding and office or finance tasks on a 1M-token context with images, for teams already on Tencent Cloud or happy to verify an account there.", "slug": "hy4", "watchFor": "Weights, licence and system card are promised but not published; tencent/Hy4 didn't answer on Hugging Face on 11 November" }, { "aheadOn": [ "Payments \u0026 pricing, 75 against 45" ], "also": [ "No key needed to call it" ], "goodFor": "A team with a four-GPU Blackwell node, or two to four Mac Studios, that wants a 1M-token multimodal coding and agent model on its own hardware with no prompts leaving it.", "slug": "muse-spark-open", "watchFor": "The NVFP4 checkpoint (about 537 GB) serves on four B300 GPUs and the BF16 one (about 1.65 TB) on eight" } ], "history": { "hy4": [ { "causes": [ "method" ], "date": "2026-11-13", "delta": 1.8, "from": 62.2, "fromGrade": "B", "to": 64, "toGrade": "B" } ], "muse-spark-open": [ { "causes": [ "method" ], "date": "2026-11-13", "delta": 3.1, "from": 53.4, "fromGrade": "D", "to": 56.5, "toGrade": "C" }, { "causes": [ "method" ], "date": "2026-10-30", "delta": -0.1, "from": 53.5, "fromGrade": "D", "to": 53.4, "toGrade": "D" }, { "causes": [ "evidence" ], "date": "2026-10-28", "delta": 0, "from": 53.5, "fromGrade": "D", "to": 53.5, "toGrade": "D" } ] }, "job": { "capability": "inference.llm", "name": "LLM inference" }, "others": [ { "json": "https://www.anchorterminal.com/compare/amazon-bedrock-vs-hy4.json", "title": "Amazon Bedrock vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/amazon-bedrock-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/amazon-bedrock-vs-muse-spark-open.json", "title": "Amazon Bedrock vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/amazon-bedrock-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/anthropic-api-vs-hy4.json", "title": "Claude API vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/anthropic-api-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/anthropic-api-vs-muse-spark-open.json", "title": "Claude API vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/anthropic-api-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/antseed-vs-hy4.json", "title": "Antseed vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/antseed-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/antseed-vs-muse-spark-open.json", "title": "Antseed vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/antseed-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/azure-openai-vs-hy4.json", "title": "Azure OpenAI in Foundry Models vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/azure-openai-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/azure-openai-vs-muse-spark-open.json", "title": "Azure OpenAI in Foundry Models vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/azure-openai-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/blockrun-ai-vs-hy4.json", "title": "BlockRun.AI vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/blockrun-ai-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/blockrun-ai-vs-muse-spark-open.json", "title": "BlockRun.AI vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/blockrun-ai-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/cerebras-inference-vs-hy4.json", "title": "Cerebras Inference vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/cerebras-inference-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/cerebras-inference-vs-muse-spark-open.json", "title": "Cerebras Inference vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/cerebras-inference-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/claude-fable-5-2-vs-hy4.json", "title": "Claude Fable 5.2 vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/claude-fable-5-2-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/claude-fable-5-2-vs-muse-spark-open.json", "title": "Claude Fable 5.2 vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/claude-fable-5-2-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/claude-mythos-5-2-vs-hy4.json", "title": "Claude Mythos 5.2 vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/claude-mythos-5-2-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/claude-mythos-5-2-vs-muse-spark-open.json", "title": "Claude Mythos 5.2 vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/claude-mythos-5-2-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/cloudflare-ai-gateway-vs-hy4.json", "title": "Cloudflare AI Gateway vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/cloudflare-ai-gateway-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/cloudflare-ai-gateway-vs-muse-spark-open.json", "title": "Cloudflare AI Gateway vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/cloudflare-ai-gateway-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/cloudflare-workers-ai-vs-hy4.json", "title": "Cloudflare Workers AI vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/cloudflare-workers-ai-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/cloudflare-workers-ai-vs-muse-spark-open.json", "title": "Cloudflare Workers AI vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/cloudflare-workers-ai-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/cohere-chat-vs-hy4.json", "title": "Cohere Chat API (Command models) vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/cohere-chat-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/cohere-chat-vs-muse-spark-open.json", "title": "Cohere Chat API (Command models) vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/cohere-chat-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/deepinfra-vs-hy4.json", "title": "DeepInfra vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/deepinfra-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/deepinfra-vs-muse-spark-open.json", "title": "DeepInfra vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/deepinfra-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/deepseek-api-vs-hy4.json", "title": "DeepSeek API vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/deepseek-api-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/deepseek-api-vs-muse-spark-open.json", "title": "DeepSeek API vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/deepseek-api-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/eden-ai-vs-hy4.json", "title": "Eden AI vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/eden-ai-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/eden-ai-vs-muse-spark-open.json", "title": "Eden AI vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/eden-ai-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/elorian-glimpse-vs-hy4.json", "title": "Glimpse vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/elorian-glimpse-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/elorian-glimpse-vs-muse-spark-open.json", "title": "Glimpse vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/elorian-glimpse-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/fireworks-inference-vs-hy4.json", "title": "Fireworks AI Inference vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/fireworks-inference-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/fireworks-inference-vs-muse-spark-open.json", "title": "Fireworks AI Inference vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/fireworks-inference-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/gemini-4-carbon-vs-hy4.json", "title": "Gemini 4 Carbon vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/gemini-4-carbon-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/gemini-4-carbon-vs-muse-spark-open.json", "title": "Gemini 4 Carbon vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/gemini-4-carbon-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/gemini-4-hydrogen-vs-hy4.json", "title": "Gemini 4 Hydrogen vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/gemini-4-hydrogen-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/gemini-4-hydrogen-vs-muse-spark-open.json", "title": "Gemini 4 Hydrogen vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/gemini-4-hydrogen-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/gemini-api-vs-hy4.json", "title": "Gemini Developer API vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/gemini-api-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/gemini-api-vs-muse-spark-open.json", "title": "Gemini Developer API vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/gemini-api-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/groq-vs-hy4.json", "title": "GroqCloud vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/groq-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/groq-vs-muse-spark-open.json", "title": "GroqCloud vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/groq-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/hugging-face-inference-providers-vs-hy4.json", "title": "Hugging Face Inference Providers vs Tencent Hy4", "url": "https://www.anchorterminal.com/compare/hugging-face-inference-providers-vs-hy4" }, { "json": "https://www.anchorterminal.com/compare/hugging-face-inference-providers-vs-muse-spark-open.json", "title": "Hugging Face Inference Providers vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/hugging-face-inference-providers-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-jais-3.json", "title": "Tencent Hy4 vs Jais 3", "url": "https://www.anchorterminal.com/compare/hy4-vs-jais-3" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-meta-model-api.json", "title": "Tencent Hy4 vs Meta Model API", "url": "https://www.anchorterminal.com/compare/hy4-vs-meta-model-api" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-minimax-api.json", "title": "Tencent Hy4 vs MiniMax API", "url": "https://www.anchorterminal.com/compare/hy4-vs-minimax-api" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-mistral-api.json", "title": "Tencent Hy4 vs Mistral AI API", "url": "https://www.anchorterminal.com/compare/hy4-vs-mistral-api" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-muse-ember.json", "title": "Tencent Hy4 vs Muse Ember", "url": "https://www.anchorterminal.com/compare/hy4-vs-muse-ember" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-nemotron-3-5-super-vl.json", "title": "Tencent Hy4 vs Nemotron 3.5 Super VL", "url": "https://www.anchorterminal.com/compare/hy4-vs-nemotron-3-5-super-vl" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-novita-ai.json", "title": "Tencent Hy4 vs Novita AI", "url": "https://www.anchorterminal.com/compare/hy4-vs-novita-ai" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-openai-api.json", "title": "Tencent Hy4 vs OpenAI API", "url": "https://www.anchorterminal.com/compare/hy4-vs-openai-api" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-openrouter.json", "title": "Tencent Hy4 vs OpenRouter", "url": "https://www.anchorterminal.com/compare/hy4-vs-openrouter" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-prism-inference.json", "title": "Tencent Hy4 vs Prism Inference", "url": "https://www.anchorterminal.com/compare/hy4-vs-prism-inference" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-ramp-router.json", "title": "Tencent Hy4 vs Router.com (Ramp Router)", "url": "https://www.anchorterminal.com/compare/hy4-vs-ramp-router" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-reflection-beam.json", "title": "Tencent Hy4 vs Beam 1.0", "url": "https://www.anchorterminal.com/compare/hy4-vs-reflection-beam" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-routerplus.json", "title": "Tencent Hy4 vs RouterPlus", "url": "https://www.anchorterminal.com/compare/hy4-vs-routerplus" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-sambanova.json", "title": "Tencent Hy4 vs SambaCloud", "url": "https://www.anchorterminal.com/compare/hy4-vs-sambanova" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-siliconflow.json", "title": "Tencent Hy4 vs SiliconFlow", "url": "https://www.anchorterminal.com/compare/hy4-vs-siliconflow" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-simplismart.json", "title": "Tencent Hy4 vs Simplismart", "url": "https://www.anchorterminal.com/compare/hy4-vs-simplismart" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-stepfun-api.json", "title": "Tencent Hy4 vs StepFun API", "url": "https://www.anchorterminal.com/compare/hy4-vs-stepfun-api" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-swe-2-flash.json", "title": "Tencent Hy4 vs SWE-2 Flash", "url": "https://www.anchorterminal.com/compare/hy4-vs-swe-2-flash" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-tencent-tokenhub.json", "title": "Tencent Hy4 vs Tencent Cloud TokenHub", "url": "https://www.anchorterminal.com/compare/hy4-vs-tencent-tokenhub" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-together-inference.json", "title": "Tencent Hy4 vs Together AI Inference", "url": "https://www.anchorterminal.com/compare/hy4-vs-together-inference" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-vercel-ai-gateway.json", "title": "Tencent Hy4 vs Vercel AI Gateway", "url": "https://www.anchorterminal.com/compare/hy4-vs-vercel-ai-gateway" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-vertex-ai.json", "title": "Tencent Hy4 vs Vertex AI (Gemini Enterprise Agent Platform)", "url": "https://www.anchorterminal.com/compare/hy4-vs-vertex-ai" }, { "json": "https://www.anchorterminal.com/compare/hy4-vs-xai-api.json", "title": "Tencent Hy4 vs xAI API (Grok)", "url": "https://www.anchorterminal.com/compare/hy4-vs-xai-api" }, { "json": "https://www.anchorterminal.com/compare/jais-3-vs-muse-spark-open.json", "title": "Jais 3 vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/jais-3-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/meta-model-api-vs-muse-spark-open.json", "title": "Meta Model API vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/meta-model-api-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/minimax-api-vs-muse-spark-open.json", "title": "MiniMax API vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/minimax-api-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/mistral-api-vs-muse-spark-open.json", "title": "Mistral AI API vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/mistral-api-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/muse-ember-vs-muse-spark-open.json", "title": "Muse Ember vs Muse Spark Open", "url": "https://www.anchorterminal.com/compare/muse-ember-vs-muse-spark-open" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-nemotron-3-5-super-vl.json", "title": "Muse Spark Open vs Nemotron 3.5 Super VL", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-nemotron-3-5-super-vl" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-novita-ai.json", "title": "Muse Spark Open vs Novita AI", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-novita-ai" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-openai-api.json", "title": "Muse Spark Open vs OpenAI API", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-openai-api" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-openrouter.json", "title": "Muse Spark Open vs OpenRouter", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-openrouter" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-prism-inference.json", "title": "Muse Spark Open vs Prism Inference", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-prism-inference" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-ramp-router.json", "title": "Muse Spark Open vs Router.com (Ramp Router)", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-ramp-router" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-reflection-beam.json", "title": "Muse Spark Open vs Beam 1.0", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-reflection-beam" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-routerplus.json", "title": "Muse Spark Open vs RouterPlus", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-routerplus" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-sambanova.json", "title": "Muse Spark Open vs SambaCloud", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-sambanova" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-siliconflow.json", "title": "Muse Spark Open vs SiliconFlow", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-siliconflow" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-simplismart.json", "title": "Muse Spark Open vs Simplismart", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-simplismart" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-stepfun-api.json", "title": "Muse Spark Open vs StepFun API", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-stepfun-api" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-swe-2-flash.json", "title": "Muse Spark Open vs SWE-2 Flash", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-swe-2-flash" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-tencent-tokenhub.json", "title": "Muse Spark Open vs Tencent Cloud TokenHub", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-tencent-tokenhub" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-together-inference.json", "title": "Muse Spark Open vs Together AI Inference", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-together-inference" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-vercel-ai-gateway.json", "title": "Muse Spark Open vs Vercel AI Gateway", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-vercel-ai-gateway" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-vertex-ai.json", "title": "Muse Spark Open vs Vertex AI (Gemini Enterprise Agent Platform)", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-vertex-ai" }, { "json": "https://www.anchorterminal.com/compare/muse-spark-open-vs-xai-api.json", "title": "Muse Spark Open vs xAI API (Grok)", "url": "https://www.anchorterminal.com/compare/muse-spark-open-vs-xai-api" } ], "panel": null, "scores": [ { "by": 4, "edge": "hy4", "hy4": 46, "key": "reliability", "muse-spark-open": 42, "name": "Reliability", "weight": 16 }, { "key": "performance", "name": "Performance", "pending": true, "weight": 10 }, { "by": 12, "edge": "hy4", "hy4": 73, "key": "schema", "muse-spark-open": 61, "name": "Schema \u0026 documentation", "weight": 13 }, { "by": 17, "edge": "hy4", "hy4": 86, "key": "ergonomics", "muse-spark-open": 69, "name": "Agent ergonomics", "weight": 13 }, { "by": 27, "edge": "hy4", "hy4": 71, "key": "security", "muse-spark-open": 44, "name": "Security \u0026 auth", "weight": 14 }, { "by": 30, "edge": "muse-spark-open", "hy4": 45, "key": "payments", "muse-spark-open": 75, "name": "Payments \u0026 pricing", "weight": 10 }, { "key": "tasks", "name": "Task success", "pending": true, "weight": 10 }, { "by": 13, "edge": "hy4", "hy4": 73, "key": "maintenance", "muse-spark-open": 60, "name": "Maintenance \u0026 community", "weight": 7 }, { "by": 1, "edge": "muse-spark-open", "hy4": 52, "key": "transparency", "muse-spark-open": 53, "name": "Transparency \u0026 trust", "weight": 7 } ], "summary": "Tencent Hy4 scores 64 (B) on agent readiness against Muse Spark Open's 56.5 (C), and leads in 5 of 7 scored categories. Muse Spark Open leads on payments \u0026 pricing. Both do llm inference.", "tasks": null, "tutorials": [], "verdicts": { "hy4": "Tencent's flagship model is on TokenHub as hy4, with a 1M-token context, image understanding, function calling and caching, at ¥6 input and ¥18 output per million tokens, half price until the end of 2026. The weights it is called open source for aren't published yet, and every benchmark is Tencent's own.", "muse-spark-open": "Muse Spark Open runs on the owner's hardware in seven runtimes, with OpenAI-compatible servers in vLLM and SGLang, standard tool calls, a 1M-token context and Meta's own serving docs. Serving the 537 GB NVFP4 checkpoint takes four B300 GPUs, Meta's download is gated by manual review, and the licence adds attribution and a revenue threshold for model resellers." } }, "kind": "anchor.page", "links": { "api": "https://www.anchorterminal.com/api/v1/index.json", "html": "https://www.anchorterminal.com/compare/hy4-vs-muse-spark-open", "json": "https://www.anchorterminal.com/compare/hy4-vs-muse-spark-open.json", "llms": "https://www.anchorterminal.com/llms.txt", "markdown": "https://www.anchorterminal.com/compare/hy4-vs-muse-spark-open.md", "mcp": "https://www.anchorterminal.com/mcp", "slim": "https://www.anchorterminal.com/compare/hy4-vs-muse-spark-open.min.md" }, "markdown": "Tencent Hy4 scores 64 (B) on agent readiness against Muse Spark Open's 56.5 (C), and leads in 5 of 7 scored categories. Muse Spark Open leads on payments \u0026 pricing. Both do llm inference.\n\n- Tencent Hy4: grade B, 64/100, rank #595 of 1524. Markdown https://www.anchorterminal.com/tools/hy4.md · JSON https://www.anchorterminal.com/api/v1/tools/hy4.json\n- Muse Spark Open: grade C, 56.5/100, rank #919 of 1524. Markdown https://www.anchorterminal.com/tools/muse-spark-open.md · JSON https://www.anchorterminal.com/api/v1/tools/muse-spark-open.json\n- Best model APIs and inference for AI agents: https://www.anchorterminal.com/best/inference/index.md\n- All 1044 models comparisons: https://www.anchorterminal.com/compare/inference/index.md\n\n## Which one, for what\n\n### Tencent Hy4 (B)\n\nGood for: Long-context agent work, coding and office or finance tasks on a 1M-token context with images, for teams already on Tencent Cloud or happy to verify an account there.\n\nAhead on:\n- Schema \u0026 documentation, 73 against 61\n- Agent ergonomics, 86 against 69\n- Security \u0026 auth, 71 against 44\n- Maintenance \u0026 community, 73 against 60\n\nWatch for: Weights, licence and system card are promised but not published; tencent/Hy4 didn't answer on Hugging Face on 11 November\n\n### Muse Spark Open (C)\n\nGood for: A team with a four-GPU Blackwell node, or two to four Mac Studios, that wants a 1M-token multimodal coding and agent model on its own hardware with no prompts leaving it.\n\nAhead on:\n- Payments \u0026 pricing, 75 against 45\n\nAlso in its favour:\n- No key needed to call it\n\nWatch for: The NVFP4 checkpoint (about 537 GB) serves on four B300 GPUs and the BF16 one (about 1.65 TB) on eight\n\n\n## Why the edge\n\n### Payments \u0026 pricing: Muse Spark Open +30\n\n- Tencent Hy4 (45): Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. No machine payment protocol (0). The price page, updated 9 November, lists ¥6 input, ¥0.3 cached and ¥18 output per million tokens, with Hy4 billed at 50% until 31 December 2026 (20). TokenHub's campaign of 1M free tokens per language model, read on 1 November, isn't confirmed for hy4 on the price page, so half (10 of 20, a judgement call). A person signs up with real-name verification (0). 0 + 20 + 10 + 0 = 30. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 30 is 45.\n- Muse Spark Open (75): Self-hosted rule. Nothing to buy from Meta, and the licence grants use free of charge. The Model as a Service clause asks for a separate agreement only from licensee groups above $20M gross revenue in 12 months (20). Free to run with no card (20). Meta's own weights need a manually reviewed Hugging Face form with legal name and date of birth, or, since the 4 November re-read of Get the model, a signed-in download from Meta's own page, so an agent can't get them alone. The ungated GGUF quantisations from ggml-org and Unsloth that Meta's docs point to download without approval (the Unsloth repository is ungated on 4 November), so part of the line (10 of 20). No payment protocol (0). Hosted copies at $1.25 input and $4.25 output per million tokens need each provider's own account and are not graded. 20 + 20 + 10 + 0 = 50. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 50 is 75.\n\n### Security \u0026 auth: Tencent Hy4 +27\n\n- Tencent Hy4 (71): Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. Bearer keys limited by model, service, IP range and token quota, long-lived with no rotation (26 of 30). The terms say inputs aren't used to train Tencent Cloud's own models, and Hy4 is first party, so no third party's terms apply (14 of 20). Retention only as long as the service needs, no periods (7 of 15). Per-call audit logs and optional full logs to the customer's CLS (14 of 15). No security.txt; Tencent's TSRC bounty covers Tencent Cloud; no certification checked (10 of 20). 26 + 14 + 7 + 14 + 10 = 71.\n- Muse Spark Open (44): Read as software the owner runs. No account or key to run it. Meta's weights need manual approval on Hugging Face and an access token, and server authentication is the owner's to add (12 of 30). An agent model that calls whatever tools the operator gives it. The card recommends human-in-the-loop confirmation for irreversible actions and Meta reports agentic-risk evaluations, but nothing in the model or template enforces an approval (8 of 20). The chat template neutralises special-token sequences and media placeholders in caller text and wraps tool results in `\u003ctool_output\u003e`, and the card advises protecting internal assistant messages and layering defences against adversarial input. No indirect prompt-injection results are published (10 of 15). No call log of its own. Visibility depends on the serving stack (5 of 15). Weights are safetensors that load through Transformers' own classes from 5.19, and Meta published cyber, chemical, biological and misalignment evaluations. dev.meta.ai/.well-known/security.txt answers 404 (re-read on 4 November), research.meta.ai per the 15 October check, and no security route is on the card or in the docs, only museoss@meta.com and the Hugging Face discussions (9 of 20). 12 + 8 + 10 + 5 + 9 = 44.\n\n### Agent ergonomics: Tencent Hy4 +17\n\n- Tencent Hy4 (86): Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. Function calling is marked for hy4 in the model list (18 of 20). Structured output marked (13 of 15). Cache-hit price published, ¥0.3 against ¥6 input (14 of 15). 1M-token context with 960K input and 64K output (15). Batch at half price covered Hy3 and Hy4 preview on 1 November; whether it covers hy4 wasn't re-read (3 of 10). OpenAI and Anthropic SDKs with a changed base URL, plus Tencent Cloud's SDK and CLI for the control plane (8 of 10). Errors with code, messages, upstream status and request ID (15 of 15). 18 + 13 + 14 + 15 + 3 + 8 + 15 = 86.\n- Muse Spark Open (69): Model reading of the checklist (tool use, structured output, caching, context, throughput, defaults and SDKs, errors). OpenAI-format tools and parallel calls come back as `tool_calls`, and the template supports namespaced tools and deferred loading through a `tool_search` tool. A forced `tool_choice` is not documented (17 of 20). JSON mode and JSON Schema come from each runtime's guided decoding, which Meta's docs show for TensorRT-LLM only (8 of 15). Prefix caching is the runtime's. Meta's SGLang command turns on hierarchical caching, and Dynamo routes each request to the worker holding its cached context (8 of 10). 1,048,576-token context, with sliding-window attention on 46 of 62 layers holding KV memory down (15 of 15). Draft models for speculative decoding in SGLang and vLLM, and no batch mode of its own (7 of 10). Any OpenAI SDK works against a local server, but reasoning can't be turned off, defaults to the most expensive level, and needs streaming, a `max_tokens` budget for reasoning and the right stop tokens (6 of 15). The template raises named errors for an unsupported reasoning effort, tool arguments passed as a string and a tool message with no name. Server errors are the runtime's (8 of 15). 17 + 8 + 8 + 15 + 7 + 6 + 8 = 69.\n\n### Maintenance \u0026 community: Tencent Hy4 +13\n\n- Tencent Hy4 (73): Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. Added on 4 November 2026 per the release notes; the model list and price page were updated on 9 November (30). Retirement notices of 21 to 33 days, no minimum written (7 of 12). Heavy retirement churn on the platform (3 of 8). Dated release notes and announcements, round-the-clock support (13 of 15). CLI replies not sampled (3 of 10). Control-plane SDKs and CLI, no inference SDK of its own (12 of 15). CLI runtime stated, CI not checked (5 of 10). 30 + 7 + 3 + 13 + 3 + 12 + 5 = 73. The model is a week old, so these lines are TokenHub's habits more than Hy4's record.\n- Muse Spark Open (60): Read for an open-weight model released three weeks ago. Released on 15 October 2026 (30). One release, BF16 and NVFP4 on the same day, with no history of revisions, successors or notice periods, and Meta Model API has no deprecation policy either (6 of 20). None of the 29 replies counted by script on 4 November across the BF16 repository's discussions came from Meta. Feedback goes to the discussions or museoss@meta.com, and Meta's docs and developer blog carry dated posts (8 of 25). Day-one support in vLLM, SGLang, TensorRT-LLM, Transformers, llama.cpp, MLX and Unsloth with Meta's own recipes, though SGLang's is a dev image and TensorRT-LLM's a feature branch (12 of 15). Current dependencies (Transformers 5.19), but no CI visible to us (4 of 10). 30 + 6 + 8 + 12 + 4 = 60.\n\n### Schema \u0026 documentation: Tencent Hy4 +12\n\n- Tencent Hy4 (73): Model reading for hy4 on Tencent Cloud TokenHub. Platform facts are the tencent-tokenhub listing's (checked 1 November), carried over as RULES.md allows for a model on an already graded platform. No OpenAPI file for inference; field pages for the OpenAI and Anthropic formats as TokenHub implements them (12 of 25). llms.txt with Markdown copies (10). The model list, updated 9 November, gives hy4's context, input and output limits and marks thinking, structured output, function calling, caching and image understanding; computer use, which Tencent's release names, isn't marked (13 of 20). Typed field pages (10 of 15). Examples in five languages and an error page of 30 codes (14 of 15). Dated release notes, with hy4 added on 4 November (14 of 15). 12 + 10 + 13 + 10 + 14 + 14 = 73.\n- Muse Spark Open (61): Read for a model the owner serves. No specification of its own. The contract is the Jinja chat template, which renders OpenAI-format tools with `additionalProperties: false` added and rejects bad values, plus the OpenAI-compatible servers in vLLM and SGLang with the `muse` parsers (10 of 25). dev.meta.ai/llms.txt lists the Muse Spark Open overview and seven pages under it, plus three cookbook pages, each with a Markdown copy (10). The card lists seven intended uses and six limitations, the docs say when to pick BF16 or NVFP4 and which reasoning strength suits which task, and the Usage Policy states what is out of scope (17 of 20). `reasoning_effort` is an enum of four values and anything else returns an error, and tools take JSON Schema. Structured output is documented only for TensorRT-LLM's `response_format`, not in the prompting guide or for vLLM and SGLang (11 of 15). curl and Python examples for every runtime and hosted provider, the rendered prompt for each case, and the template's own error messages, but no error reference for a served model (11 of 15). One release, with no changelog or version scheme beyond the licence's version line, Muse Spark Open (2 of 15). 10 + 10 + 17 + 11 + 11 + 2 = 61.\n\n## Score by category\n\n| Category | Weight | Tencent Hy4 | Muse Spark Open | Edge |\n| --- | --- | --- | --- | --- |\n| Reliability | 16% (20 this run) | 46 | 42 | Tencent Hy4 +4 |\n| Performance | 10%, pending | pending | pending | not scored in this run |\n| Schema \u0026 documentation | 13% (16.2 this run) | 73 | 61 | Tencent Hy4 +12 |\n| Agent ergonomics | 13% (16.2 this run) | 86 | 69 | Tencent Hy4 +17 |\n| Security \u0026 auth | 14% (17.5 this run) | 71 | 44 | Tencent Hy4 +27 |\n| Payments \u0026 pricing | 10% (12.5 this run) | 45 | 75 | Muse Spark Open +30 |\n| Task success | 10%, pending | pending | pending | not scored in this run |\n| Maintenance \u0026 community | 7% (8.8 this run) | 73 | 60 | Tencent Hy4 +13 |\n| Transparency \u0026 trust | 7% (8.8 this run) | 52 | 53 | Muse Spark Open +1 |\n| Negative events | ≤15 | 0 | 0 | |\n| **Total** | | **64 · B** | **56.5 · C** | |\n\n## Facts side by side\n\n| Fact | Tencent Hy4 | Muse Spark Open |\n| --- | --- | --- |\n| Kind | Model API | Model API |\n| Vendor | Tencent (Hy team) | Meta |\n| Hosted endpoint | no (local only) | no (local only) |\n| Transports | HTTP | HTTP |\n| Auth | API key | None |\n| Pricing | Pay per use | Free |\n| x402 | no | no |\n| Licence | Hosted model under the Tencent Cloud TokenHub terms. Tencent says weights and a system card will follow; no licence is stated yet | Muse Model Licence (source-available weights, not an OSI licence), free of charge with an attribution notice, a separate agreement with Meta before Model as a Service by a licensee group above $20M gross revenue, and the Muse Spark Open Usage Policy |\n| Read-only variant documented | no | no |\n| llms.txt | yes | yes |\n| Last release | 2026-11-04 | 2026-10-15 |\n| Terms last updated | couldn't be read | 2026-10-15 |\n| Privacy policy last updated | couldn't be read | no document linked |\n| Customer content may train models | couldn't be read | not found in the text |\n| Terms restrict automated access | couldn't be read | not found in the text |\n| Terms restrict benchmarking | couldn't be read | not found in the text |\n| Terms or service can change without notice | couldn't be read | not found in the text |\n| Arbitration or class-action waiver | couldn't be read | not found in the text |\n\n## How the scores have moved\n\n### Tencent Hy4\n\n- +1.8 on 13 November 2026: checklist\n\n### Muse Spark Open\n\n- +3.1 on 13 November 2026 (D to C): checklist\n- -0.1 on 30 October 2026: checklist\n- no change on 28 October 2026: new evidence\n\n## Verdicts\n\n**Tencent Hy4.** Tencent's flagship model is on TokenHub as hy4, with a 1M-token context, image understanding, function calling and caching, at ¥6 input and ¥18 output per million tokens, half price until the end of 2026. The weights it is called open source for aren't published yet, and every benchmark is Tencent's own.\n\n**Muse Spark Open.** Muse Spark Open runs on the owner's hardware in seven runtimes, with OpenAI-compatible servers in vLLM and SGLang, standard tool calls, a 1M-token context and Meta's own serving docs. Serving the 537 GB NVFP4 checkpoint takes four B300 GPUs, Meta's download is gated by manual review, and the licence adds attribution and a revenue threshold for model resellers.\n\n## Before you call either\n\n### Tencent Hy4\n\n1. Use model id hy4 on TokenHub's OpenAI- or Anthropic-compatible endpoints; hy4-preview is a separate, older model\n2. Keep input under 960K tokens and expect at most 64K output\n3. Reuse prompt prefixes: cached input is ¥0.3 against ¥6\n4. Check the price page for the end of the 50% discount on 31 December 2026 before budgeting\n5. Read rate limits from the console or DescribeModelQuota; the public docs don't give them for chat models\n\n### Muse Spark Open\n\n1. Start vLLM or SGLang with `--reasoning-parser muse --tool-call-parser muse` (vLLM also `--enable-auto-tool-choice`), or tool calls come back as raw text\n2. Set `reasoning_effort` to `low` or `medium` for short tasks. The default is `xhigh`, reasoning can't be turned off, and other values return an error\n3. Stop only on `\u003c|eot|\u003e` and end of text. Stopping on `\u003c|eom|\u003e` cuts off tool calls and final answers\n4. Stream long generations and leave `max_tokens` room for reasoning, or `content` can come back empty with `finish_reason` `length`\n5. Don't call Meta Model API for this model. Use a self-hosted server or a hosted provider's model string, such as `meta-models/muse-spark-open`\n\n## Questions\n\n### Which is better for AI agents, Tencent Hy4 or Muse Spark Open?\n\nTencent Hy4 scores 64 (B) on agent readiness against Muse Spark Open's 56.5 (C), and leads in 5 of 7 scored categories. Muse Spark Open leads on payments \u0026 pricing.\n\n### Do Tencent Hy4 and Muse Spark Open need an API key?\n\nTencent Hy4 needs an API key. Muse Spark Open needs no key.\n\n### Can an agent call Tencent Hy4 and Muse Spark Open without installing anything?\n\nNo hosted endpoint is listed for Tencent Hy4. No hosted endpoint is listed for Muse Spark Open.\n\n\n## For agents\n\n- This comparison as JSON: https://www.anchorterminal.com/compare/hy4-vs-muse-spark-open.json, and with the fewest tokens: https://www.anchorterminal.com/compare/hy4-vs-muse-spark-open.min.md\n- Over MCP at https://www.anchorterminal.com/mcp (no key): `compare_tools {\"a\": \"hy4\", \"b\": \"muse-spark-open\"}`. From a terminal: `anchor compare hy4 muse-spark-open`\n- Each listing in full: https://www.anchorterminal.com/api/v1/tools/hy4.json and https://www.anchorterminal.com/api/v1/tools/muse-spark-open.json\n\n## Other comparisons with Tencent Hy4 or Muse Spark Open\n\n- [Amazon Bedrock vs Tencent Hy4](https://www.anchorterminal.com/compare/amazon-bedrock-vs-hy4.md)\n- [Amazon Bedrock vs Muse Spark Open](https://www.anchorterminal.com/compare/amazon-bedrock-vs-muse-spark-open.md)\n- [Claude API vs Tencent Hy4](https://www.anchorterminal.com/compare/anthropic-api-vs-hy4.md)\n- [Claude API vs Muse Spark Open](https://www.anchorterminal.com/compare/anthropic-api-vs-muse-spark-open.md)\n- [Antseed vs Tencent Hy4](https://www.anchorterminal.com/compare/antseed-vs-hy4.md)\n- [Antseed vs Muse Spark Open](https://www.anchorterminal.com/compare/antseed-vs-muse-spark-open.md)\n- [Azure OpenAI in Foundry Models vs Tencent Hy4](https://www.anchorterminal.com/compare/azure-openai-vs-hy4.md)\n- [Azure OpenAI in Foundry Models vs Muse Spark Open](https://www.anchorterminal.com/compare/azure-openai-vs-muse-spark-open.md)\n- [BlockRun.AI vs Tencent Hy4](https://www.anchorterminal.com/compare/blockrun-ai-vs-hy4.md)\n- [BlockRun.AI vs Muse Spark Open](https://www.anchorterminal.com/compare/blockrun-ai-vs-muse-spark-open.md)\n- [Cerebras Inference vs Tencent Hy4](https://www.anchorterminal.com/compare/cerebras-inference-vs-hy4.md)\n- [Cerebras Inference vs Muse Spark Open](https://www.anchorterminal.com/compare/cerebras-inference-vs-muse-spark-open.md)\n- [Claude Fable 5.2 vs Tencent Hy4](https://www.anchorterminal.com/compare/claude-fable-5-2-vs-hy4.md)\n- [Claude Fable 5.2 vs Muse Spark Open](https://www.anchorterminal.com/compare/claude-fable-5-2-vs-muse-spark-open.md)\n- [Claude Mythos 5.2 vs Tencent Hy4](https://www.anchorterminal.com/compare/claude-mythos-5-2-vs-hy4.md)\n- [Claude Mythos 5.2 vs Muse Spark Open](https://www.anchorterminal.com/compare/claude-mythos-5-2-vs-muse-spark-open.md)\n- [Cloudflare AI Gateway vs Tencent Hy4](https://www.anchorterminal.com/compare/cloudflare-ai-gateway-vs-hy4.md)\n- [Cloudflare AI Gateway vs Muse Spark Open](https://www.anchorterminal.com/compare/cloudflare-ai-gateway-vs-muse-spark-open.md)\n- [Cloudflare Workers AI vs Tencent Hy4](https://www.anchorterminal.com/compare/cloudflare-workers-ai-vs-hy4.md)\n- [Cloudflare Workers AI vs Muse Spark Open](https://www.anchorterminal.com/compare/cloudflare-workers-ai-vs-muse-spark-open.md)\n- [Cohere Chat API (Command models) vs Tencent Hy4](https://www.anchorterminal.com/compare/cohere-chat-vs-hy4.md)\n- [Cohere Chat API (Command models) vs Muse Spark Open](https://www.anchorterminal.com/compare/cohere-chat-vs-muse-spark-open.md)\n- [DeepInfra vs Tencent Hy4](https://www.anchorterminal.com/compare/deepinfra-vs-hy4.md)\n- [DeepInfra vs Muse Spark Open](https://www.anchorterminal.com/compare/deepinfra-vs-muse-spark-open.md)\n- [DeepSeek API vs Tencent Hy4](https://www.anchorterminal.com/compare/deepseek-api-vs-hy4.md)\n- [DeepSeek API vs Muse Spark Open](https://www.anchorterminal.com/compare/deepseek-api-vs-muse-spark-open.md)\n- [Eden AI vs Tencent Hy4](https://www.anchorterminal.com/compare/eden-ai-vs-hy4.md)\n- [Eden AI vs Muse Spark Open](https://www.anchorterminal.com/compare/eden-ai-vs-muse-spark-open.md)\n- [Glimpse vs Tencent Hy4](https://www.anchorterminal.com/compare/elorian-glimpse-vs-hy4.md)\n- [Glimpse vs Muse Spark Open](https://www.anchorterminal.com/compare/elorian-glimpse-vs-muse-spark-open.md)\n- [Fireworks AI Inference vs Tencent Hy4](https://www.anchorterminal.com/compare/fireworks-inference-vs-hy4.md)\n- [Fireworks AI Inference vs Muse Spark Open](https://www.anchorterminal.com/compare/fireworks-inference-vs-muse-spark-open.md)\n- [Gemini 4 Carbon vs Tencent Hy4](https://www.anchorterminal.com/compare/gemini-4-carbon-vs-hy4.md)\n- [Gemini 4 Carbon vs Muse Spark Open](https://www.anchorterminal.com/compare/gemini-4-carbon-vs-muse-spark-open.md)\n- [Gemini 4 Hydrogen vs Tencent Hy4](https://www.anchorterminal.com/compare/gemini-4-hydrogen-vs-hy4.md)\n- [Gemini 4 Hydrogen vs Muse Spark Open](https://www.anchorterminal.com/compare/gemini-4-hydrogen-vs-muse-spark-open.md)\n- [Gemini Developer API vs Tencent Hy4](https://www.anchorterminal.com/compare/gemini-api-vs-hy4.md)\n- [Gemini Developer API vs Muse Spark Open](https://www.anchorterminal.com/compare/gemini-api-vs-muse-spark-open.md)\n- [GroqCloud vs Tencent Hy4](https://www.anchorterminal.com/compare/groq-vs-hy4.md)\n- [GroqCloud vs Muse Spark Open](https://www.anchorterminal.com/compare/groq-vs-muse-spark-open.md)\n- [Hugging Face Inference Providers vs Tencent Hy4](https://www.anchorterminal.com/compare/hugging-face-inference-providers-vs-hy4.md)\n- [Hugging Face Inference Providers vs Muse Spark Open](https://www.anchorterminal.com/compare/hugging-face-inference-providers-vs-muse-spark-open.md)\n- [Tencent Hy4 vs Jais 3](https://www.anchorterminal.com/compare/hy4-vs-jais-3.md)\n- [Tencent Hy4 vs Meta Model API](https://www.anchorterminal.com/compare/hy4-vs-meta-model-api.md)\n- [Tencent Hy4 vs MiniMax API](https://www.anchorterminal.com/compare/hy4-vs-minimax-api.md)\n- [Tencent Hy4 vs Mistral AI API](https://www.anchorterminal.com/compare/hy4-vs-mistral-api.md)\n- [Tencent Hy4 vs Muse Ember](https://www.anchorterminal.com/compare/hy4-vs-muse-ember.md)\n- [Tencent Hy4 vs Nemotron 3.5 Super VL](https://www.anchorterminal.com/compare/hy4-vs-nemotron-3-5-super-vl.md)\n- [Tencent Hy4 vs Novita AI](https://www.anchorterminal.com/compare/hy4-vs-novita-ai.md)\n- [Tencent Hy4 vs OpenAI API](https://www.anchorterminal.com/compare/hy4-vs-openai-api.md)\n- [Tencent Hy4 vs OpenRouter](https://www.anchorterminal.com/compare/hy4-vs-openrouter.md)\n- [Tencent Hy4 vs Prism Inference](https://www.anchorterminal.com/compare/hy4-vs-prism-inference.md)\n- [Tencent Hy4 vs Router.com (Ramp Router)](https://www.anchorterminal.com/compare/hy4-vs-ramp-router.md)\n- [Tencent Hy4 vs Beam 1.0](https://www.anchorterminal.com/compare/hy4-vs-reflection-beam.md)\n- [Tencent Hy4 vs RouterPlus](https://www.anchorterminal.com/compare/hy4-vs-routerplus.md)\n- [Tencent Hy4 vs SambaCloud](https://www.anchorterminal.com/compare/hy4-vs-sambanova.md)\n- [Tencent Hy4 vs SiliconFlow](https://www.anchorterminal.com/compare/hy4-vs-siliconflow.md)\n- [Tencent Hy4 vs Simplismart](https://www.anchorterminal.com/compare/hy4-vs-simplismart.md)\n- [Tencent Hy4 vs StepFun API](https://www.anchorterminal.com/compare/hy4-vs-stepfun-api.md)\n- [Tencent Hy4 vs SWE-2 Flash](https://www.anchorterminal.com/compare/hy4-vs-swe-2-flash.md)\n- [Tencent Hy4 vs Tencent Cloud TokenHub](https://www.anchorterminal.com/compare/hy4-vs-tencent-tokenhub.md)\n- [Tencent Hy4 vs Together AI Inference](https://www.anchorterminal.com/compare/hy4-vs-together-inference.md)\n- [Tencent Hy4 vs Vercel AI Gateway](https://www.anchorterminal.com/compare/hy4-vs-vercel-ai-gateway.md)\n- [Tencent Hy4 vs Vertex AI (Gemini Enterprise Agent Platform)](https://www.anchorterminal.com/compare/hy4-vs-vertex-ai.md)\n- [Tencent Hy4 vs xAI API (Grok)](https://www.anchorterminal.com/compare/hy4-vs-xai-api.md)\n- [Jais 3 vs Muse Spark Open](https://www.anchorterminal.com/compare/jais-3-vs-muse-spark-open.md)\n- [Meta Model API vs Muse Spark Open](https://www.anchorterminal.com/compare/meta-model-api-vs-muse-spark-open.md)\n- [MiniMax API vs Muse Spark Open](https://www.anchorterminal.com/compare/minimax-api-vs-muse-spark-open.md)\n- [Mistral AI API vs Muse Spark Open](https://www.anchorterminal.com/compare/mistral-api-vs-muse-spark-open.md)\n- [Muse Ember vs Muse Spark Open](https://www.anchorterminal.com/compare/muse-ember-vs-muse-spark-open.md)\n- [Muse Spark Open vs Nemotron 3.5 Super VL](https://www.anchorterminal.com/compare/muse-spark-open-vs-nemotron-3-5-super-vl.md)\n- [Muse Spark Open vs Novita AI](https://www.anchorterminal.com/compare/muse-spark-open-vs-novita-ai.md)\n- [Muse Spark Open vs OpenAI API](https://www.anchorterminal.com/compare/muse-spark-open-vs-openai-api.md)\n- [Muse Spark Open vs OpenRouter](https://www.anchorterminal.com/compare/muse-spark-open-vs-openrouter.md)\n- [Muse Spark Open vs Prism Inference](https://www.anchorterminal.com/compare/muse-spark-open-vs-prism-inference.md)\n- [Muse Spark Open vs Router.com (Ramp Router)](https://www.anchorterminal.com/compare/muse-spark-open-vs-ramp-router.md)\n- [Muse Spark Open vs Beam 1.0](https://www.anchorterminal.com/compare/muse-spark-open-vs-reflection-beam.md)\n- [Muse Spark Open vs RouterPlus](https://www.anchorterminal.com/compare/muse-spark-open-vs-routerplus.md)\n- [Muse Spark Open vs SambaCloud](https://www.anchorterminal.com/compare/muse-spark-open-vs-sambanova.md)\n- [Muse Spark Open vs SiliconFlow](https://www.anchorterminal.com/compare/muse-spark-open-vs-siliconflow.md)\n- [Muse Spark Open vs Simplismart](https://www.anchorterminal.com/compare/muse-spark-open-vs-simplismart.md)\n- [Muse Spark Open vs StepFun API](https://www.anchorterminal.com/compare/muse-spark-open-vs-stepfun-api.md)\n- [Muse Spark Open vs SWE-2 Flash](https://www.anchorterminal.com/compare/muse-spark-open-vs-swe-2-flash.md)\n- [Muse Spark Open vs Tencent Cloud TokenHub](https://www.anchorterminal.com/compare/muse-spark-open-vs-tencent-tokenhub.md)\n- [Muse Spark Open vs Together AI Inference](https://www.anchorterminal.com/compare/muse-spark-open-vs-together-inference.md)\n- [Muse Spark Open vs Vercel AI Gateway](https://www.anchorterminal.com/compare/muse-spark-open-vs-vercel-ai-gateway.md)\n- [Muse Spark Open vs Vertex AI (Gemini Enterprise Agent Platform)](https://www.anchorterminal.com/compare/muse-spark-open-vs-vertex-ai.md)\n- [Muse Spark Open vs xAI API (Grok)](https://www.anchorterminal.com/compare/muse-spark-open-vs-xai-api.md)\n", "meta": { "attribution": "Anchor Terminal (https://www.anchorterminal.com)", "attributionRequired": true, "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)", "docs": "https://www.anchorterminal.com/docs/", "generatedAt": "2026-11-14", "license": "CC-BY-4.0", "method": "https://www.anchorterminal.com/benchmark/", "methodology": "0.8", "openapi": "https://www.anchorterminal.com/openapi.json", "preview": false, "run": "2026-10-01", "runLabel": "October 2026 research run", "terms": "https://www.anchorterminal.com/terms/#data" }, "page": { "breadcrumbs": [ { "name": "Home", "url": "https://www.anchorterminal.com/" }, { "name": "Compare", "url": "https://www.anchorterminal.com/compare/" }, { "name": "Tencent Hy4 vs Muse Spark Open", "url": "" } ], "description": "Tencent Hy4 scores 64 (B) to Muse Spark Open's 56.5 (C) for llm inference. Prices, MCP, x402, uptime and agent notes side by side.", "facts": [ "Tencent Hy4 B 64", "Muse Spark Open C 56.5", "scores" ], "h1": "Tencent Hy4 vs Muse Spark Open", "image": "https://www.anchorterminal.com/assets/og/default.png", "path": "/compare/hy4-vs-muse-spark-open", "published": "2026-10-01", "section": "tools", "title": "Tencent Hy4 vs Muse Spark Open for AI agents (2026) | Anchor Terminal", "toc": null, "updated": "2026-11-13", "url": "https://www.anchorterminal.com/compare/hy4-vs-muse-spark-open" }, "tokens": { "markdown": 6750, "slim": 980 }, "version": 1 }