Jais 3 vs Muse Ember

Jais 3 60

Self-hosted rule. Apache-2.0 weights with nothing to buy and no hosted option from the makers (20). Free to run with no card (20). Downloads are ungated and anonymous, so an agent can fetch the weights with no sign-up (20). No payment protocol (0). 20 + 20 + 20 + 0 = 60.

Muse Ember 20

Published rubric. No payment protocol (0). Per-token prices for both tiers published without a login, read 9 November (20). No free API allowance found; billing needs a card, per the platform reading (0). Browser sign-up (0). 0 + 20 + 0 + 0 = 20.

Jais 3 31

Read as software the owner runs. No account or key, ungated downloads, and server authentication is the owner's to add (12 of 30). A model that calls whatever tools the operator wires in. The cards rule high-stakes decisions without human oversight out of scope, but nothing in the model or template enforces approval (6 of 20). The template renders tool results in their own tool role, but string results go in as raw text (only non-string content is serialised with tojson), no escaping of special-token sequences in caller or tool text was found, and no injection guidance or results are published (4 of 15). No call log of its own. Visibility depends on the serving stack (5 of 15). Every launch command passes --trust-remote-code, so modeling_k2_horizon.py from the repository runs on the host. We read it and found only torch and transformers imports. No security.txt on ifm.ai (404 on 17 October, not re-read), no security contact on the cards and no safety evaluation found. Since 25 October the cards link a technical report on ifm.ai, which answered a Cloudflare bot check on 4 November, so what it covers is unread (4 of 20). 12 + 6 + 4 + 5 + 4 = 31.

Muse Ember 64

Platform readings are the meta-model-api listing's of 4 November. Team keys, named, revocable, no scopes (20 of 30). Standard mode only, so content isn't used to train Meta models under the terms, with safety use reserved (17 of 20). Retention 'as needed' with zero retention through sales (10 of 15). Usage dashboard by model, no per-call log (5 of 15). No security.txt on dev.meta.ai, a bug bounty and a security commitments page, no named certification (12 of 20). 20 + 17 + 10 + 5 + 12 = 64.

Jais 3 62

Read for an open-weight model released on 15 October 2026. The last change to the template or code was on 15 October, within 30 days. The cards gained a technical report link on 25 October and a citation update on 27 October, which change no file a runtime loads (30). One family release, with dated builds before it (FP8 from 29 September, MLX from 1 October, GGUF on 12 and 13 October) and template fixes on 14 and 15 October, but no deprecation notices or successor history to judge. K2 Horizon came on 1 September and Jais 2 in December 2025 (10 of 20, the model reading of this line, held). One outside question, on MoVA-36B-A4B on 26 October, had no reply on 4 November, and IFM's GitHub repositories weren't read. The template fixes match the problem IFM's MLX cards describe (7 of 25, held). Recipes in vLLM's and SGLang's own docs, K2 Horizon support in llama.cpp's tagged builds from b11454 and FP8, GGUF and MLX builds from IFM, but no native mlx-lm support (11 of 15). Current dependencies (Transformers 5.13 to 5.15, vLLM 0.30), with no CI visible (4 of 10). 30 + 10 + 7 + 11 + 4 = 62.

Muse Ember 46

Model reading. Released 9 November 2026 (30). No deprecation policy or notice period found, per the platform reading (0 of 12). Dated launches on the API, the latest Muse Wonder on 12 October and Muse Ember on 9 November (8 of 8). No Model API changelog; a help centre with tickets (8 of 15). No SDK of Meta's own (0 of 15, 0 of 10). 30 + 0 + 8 + 8 + 0 = 46.

Jais 3 48

Read with the local-software lines, since the owner serves the weights. Ungated weights on Hugging Face with stated runtimes (vLLM 0.30.0 or later, SGLang 0.5.20 or later, Transformers validated at 5.15.0, mlx-lm with shipped model code), but every launch command trusts remote code. The GGUF card still says to build llama.cpp from master or wait for a release, though llama.cpp's own tags carry K2 Horizon support from b11454 (6 October 2026) (16 of 20, held). No CI or test suite of the makers' was found, and the code repository is 'coming soon'. SGLang's cookbook reports a validation pass of the tool-call parser on all six checkpoints, and IFM's MLX cards describe conversion checks against the PyTorch code (3 of 25). No crash reports. On 4 November the six main repositories held seven discussions, six of them closed citation updates and one open MoVA-36B-A4B thread from 26 October with no reply yet. Two template fixes for OpenAI-style message history landed on 14 October, before the launch (15 of 25). No changelog or version tags, only public commit messages (2 of 15). Released as named final checkpoints with no preview label, though mlx-lm v0.32.2 still has no K2 Horizon model code of its own (12 of 15). 16 + 3 + 15 + 2 + 12 = 48.

Muse Ember 40

Model reading, for the surface it's called on, Meta Model API. No status page with readable history, per the meta-model-api listing's 4 November check (0), so no readable incident history (5). Rate limits published per team and per model, 3,000 RPM and 4,000,000 TPM for muse-ember-1.0, read 9 November (15). 429 and 503 with Retry-After, backoff guidance and rate-limit headers, per the meta-model-api listing (15). No SLA found (0). Muse Ember carries no preview label, but the API opened as a public preview in July and no general-availability notice was found, so half, as for the platform (5). 0 + 5 + 15 + 15 + 0 + 5 = 40.

Jais 3 65

Model reading of the checklist (tool use, structured output, caching, context, throughput, defaults and SDKs, errors). OpenAI-format tools come back as tool_calls in three selectable wire formats, with a multi-turn loop example. Forced tool_choice and parallel calls aren't documented (15 of 20). JSON mode and schemas only through each runtime's guided decoding, not documented by the makers (5 of 15). Prefix caching is the runtime's, and SGLang's page covers HiCache (7 of 10). 524,288 tokens on five of six models, 131,072 on the 0.9B (14 of 15). Uno adapters claim 2.15 times faster decoding for 0.9B to 32B but need IFM's own engine. SGLang lists n-gram speculative decoding and PD disaggregation. No batch mode of its own (7 of 10). Any OpenAI SDK works against a local server, and the 0.9B to 7B fit one GPU. But the cards ask for high effort and 32,768 output tokens on every request, the template inserts an identity system prompt when none is sent, and earlier assistant turns must carry a string thinking field (8 of 15). The template raises named errors for bad options, string-encoded tool arguments and malformed schemas. Server errors are the runtime's (9 of 15). 15 + 5 + 7 + 14 + 7 + 8 + 9 = 65.

Muse Ember 72

Model reading. Tool calls, parallel calls, computer use and search grounding, with tool_choice limited to auto per the platform reading (15 of 20). Strict structured output (15). Automatic prompt caching, cached input at 4% of the input price (15). A 1,048,576-token context, but requests over 200K input tokens bill at three times the rate for the whole request (12 of 15). No batch endpoint (0). No SDK of Meta's own; OpenAI and Anthropic SDKs against compatible endpoints (0). OpenAI-style errors with fixes and retryable codes (15). 15 + 15 + 15 + 12 + 0 + 0 + 15 = 72.

Jais 3 54

Read for a model the owner serves. No specification of its own. The contract is the Jinja chat template, which validates tool JSON Schemas (required fields must exist in properties) and rejects unknown options, plus the OpenAI-compatible servers in vLLM and SGLang with the k2_horizon parsers (10 of 25). The cards are Markdown on Hugging Face, and SGLang's docs have an llms.txt, but no llms.txt or Markdown docs from IFM were found and ifm.ai couldn't be read (5 of 10). Each card states intended use, four out-of-scope uses, limitations and per-size serving advice (15 of 20). reasoning_effort and tool_call_format are enums enforced with errors, and tools take JSON Schema. The 375B card describes low as no reasoning and omits medium, which the template accepts, and structured output isn't documented (10 of 15). Python examples for a reasoning call and a tool loop, commands for five runtimes, and named template errors, but no error reference for a served model (11 of 15). No changelog or version scheme. Training-stage branches and public commit history (3 of 15). 10 + 5 + 15 + 10 + 11 + 3 = 54.

Muse Ember 60

Model reading. No OpenAPI file; the Markdown reference is the contract, per the meta-model-api listing (0). llms.txt and Markdown copies of every page (10). The Muse Ember page and choose-a-model page say what to hand it and what not to (hard single-shot coding, knowledge-heavy questions, visual coding), with model ID, inputs, context, max output and effort range (18 of 20). JSON Schema tools and output, reasoning effort an enum, per the platform reading (12 of 15). Curl, Python and TypeScript examples and the platform's error guide (15). A versioned model ID and a dated launch post, with no API changelog (5 of 15). 0 + 10 + 18 + 12 + 15 + 5 = 60.