{
  "listings": [
    {
      "slug": "granite-4-2-3b",
      "name": "Granite 4.2 3B",
      "company": "International Business Machines Corporation (per IBM's naming guidance, as read for granite-guardian)",
      "grade": "C",
      "fit": "yes",
      "why": "Tool calling with OpenAI schemas is documented, enable_thinking false gives direct answers, the Q4_K_M is 2.2 GB on Ollama, and acc-26 measured 100 percent routing at 0.54 seconds mean on a 16 GB GPU.",
      "source": "https://huggingface.co/ibm-granite/granite-4.2-3b",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/granite-4-2-3b"
    },
    {
      "slug": "qwen3-5-0-8b",
      "name": "Qwen3.5-0.8B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "D",
      "fit": "yes",
      "why": "Card names qwen3_coder for vLLM and SGLang and Ollama declares tools, but nothing measures accuracy on 10 to 60 tools at this size; only the template shows the format.",
      "source": "https://huggingface.co/Qwen/Qwen3.5-0.8B",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-5-0-8b"
    },
    {
      "slug": "qwen3-5-2b",
      "name": "Qwen3.5-2B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "D",
      "fit": "yes",
      "why": "Card names qwen3_coder for vLLM and SGLang and Ollama declares tools, but nothing measures accuracy on 10 to 60 tools at this size; only the template shows the format.",
      "source": "https://huggingface.co/Qwen/Qwen3.5-2B",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-5-2b"
    },
    {
      "slug": "agents-a1-4b",
      "name": "Agents-A1-4B",
      "company": "Shanghai Artificial Intelligence Laboratory (the InternAgent team there, per the card's recommended system prompt and the template's default prompt); the GitHub LICENSE leaves the copyright line as the Apache template",
      "grade": "C",
      "fit": "caveat",
      "why": "Native tool calling with the qwen3_coder parser and a 2.7 GB GGUF fit an 8 GB GPU, but thinking defaults on, no Ollama tag exists and no test of 30 commands was found.",
      "source": "https://huggingface.co/InternScience/Agents-A1-4B",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/agents-a1-4b"
    },
    {
      "slug": "gemma-4-12b-it",
      "name": "Gemma 4 12B IT",
      "company": "Google LLC (RDAP registrant of gemma.ai as recorded for the first batch's Gemma 4 listings; the card names Google DeepMind as author)",
      "grade": "BB",
      "fit": "caveat",
      "why": "acc-24 ran Assist on an RTX 3060 Ti 8 GB: 100 percent tool-call success but 3.6 seconds median, above the 3 second target; the 3-bit quant returned empty 9 of 33 times, IQ4_XS none.",
      "source": "https://community.home-assistant.io/t/integration-tuned-for-running-assist-on-a-local-llm-in-8-gb-vram/1025725",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gemma-4-12b-it"
    },
    {
      "slug": "gemma-4-26b-a4b-it",
      "name": "Gemma 4 26B A4B IT",
      "company": "Google LLC (RDAP registrant of gemma.ai; the card names Google DeepMind as author)",
      "grade": "BB",
      "fit": "caveat",
      "why": "Google's Q4_0 is 14.4 GB and Ollama's 16 GB, which does not fit an 8 GB GPU and leaves no context on 16 GB; acc-24 chose the 12B for this reason. Tools badge present.",
      "source": "https://huggingface.co/api/models/google/gemma-4-26B-A4B-it-qat-q4_0-gguf?blobs=true",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gemma-4-26b-a4b-it"
    },
    {
      "slug": "gemma-4-e2b-it",
      "name": "Gemma 4 E2B IT",
      "company": "Google LLC (RDAP registrant of gemma.ai as recorded for the first batch's Gemma 4 listings; the card names Google DeepMind as author)",
      "grade": "C",
      "fit": "caveat",
      "why": "Fits any 8 GB GPU with Ollama tools support, but no 30-command test exists for E2B and acc-45 rates it well below E4B on every task.",
      "source": "https://ollama.com/library/gemma4/tags",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gemma-4-e2b-it"
    },
    {
      "slug": "gemma-4-e4b-it",
      "name": "Gemma 4 E4B IT",
      "company": "Google LLC (RDAP registrant of gemma.ai as recorded for the first batch's Gemma 4 listings; the card names Google DeepMind as author)",
      "grade": "B",
      "fit": "caveat",
      "why": "Fits an 8 GB GPU with Ollama tools support and acc-45 passed a six-step agent task, but no 30-command tool test exists for E4B and acc-26 found some Gemma builds without tool templates.",
      "source": "https://ollama.com/library/gemma4/tags",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gemma-4-e4b-it"
    },
    {
      "slug": "glm-4-7-flash",
      "name": "GLM-4.7-Flash",
      "company": "JINGSHENG HENGXING TECHNOLOGY PTE.LTD (operator of Z.ai per its terms of use, and registrant of z.ai)",
      "grade": "B",
      "fit": "caveat",
      "why": "acc-23 ran Home Assistant Assist on GLM-4.7 Flash through Ollama and it worked where Qwen failed; the 19 GB Ollama build exceeds an 8 to 16 GB GPU without offload, and Unsloth's Q3_K_XL is 13.8 GB.",
      "source": "https://community.home-assistant.io/t/assist-qwen3-5-35b-a3b-tool-calling/991434",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/glm-4-7-flash"
    },
    {
      "slug": "gpt-oss-20b",
      "name": "gpt-oss-20b",
      "company": "OpenAI (named in USAGE_POLICY and the model card; the LICENSE file's copyright line is the unfilled Apache template)",
      "grade": "B",
      "fit": "caveat",
      "why": "Fits a 16 GB GPU (acc-47 on an RTX 4080) but acc-47 found tool calling unreliable and long thinking too slow for Home Assistant; no off switch.",
      "source": "https://engineered.at/articles/running-openai-s-gpt-oss-locally-the-good-the-bad-and-the-loopy",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gpt-oss-20b"
    },
    {
      "slug": "lfm2-5-2-6b",
      "name": "LFM2.5-2.6B",
      "company": "Liquid AI, Inc. (the licensor named in the LFM Open License)",
      "grade": "D",
      "fit": "caveat",
      "why": "Tool use is documented and 1.6 GB fits an 8 GB GPU, but calls come back as Pythonic lists between special tokens, thinking cannot be turned off, and a discussion calls the tool-call tokens the model's biggest problem.",
      "source": "https://huggingface.co/LiquidAI/LFM2.5-2.6B/raw/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/lfm2-5-2-6b"
    },
    {
      "slug": "ling-3-0-tiny",
      "name": "Ling-3.0-tiny",
      "company": "Ant Intelligence (Hang Zhou) Technology Co., Ltd (named as provider in the training-content summary); inclusionAI, Ant Group in the site footer of inclusion-ai.org",
      "grade": "C",
      "fit": "caveat",
      "why": "Fits 8 GB with ling3 parsers in vLLM and llama.cpp --jinja, but there is no Ollama library entry, only a manual import, and looping is reported.",
      "source": "https://huggingface.co/inclusionAI/Ling-3.0-tiny/raw/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/ling-3-0-tiny"
    },
    {
      "slug": "mimo-v2-6-distill-qwen-9b",
      "name": "MiMo-V2.6 Distill Qwen 9B",
      "company": "Xiaomi (the card and report name Xiaomi MiMo; no LICENSE file names a copyright holder)",
      "grade": "C",
      "fit": "caveat",
      "why": "Fits 8 GB through Ollama's tools-badged tag, but tool-call reliability on llama.cpp is disputed in discussions and the card documents only SGLang's mimo reasoning parser.",
      "source": "https://ollama.com/library/mimo-v2.6/tags",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/mimo-v2-6-distill-qwen-9b"
    },
    {
      "slug": "minicpm-o-4-5",
      "name": "MiniCPM-o 4.5",
      "company": "OpenBMB (named as the MiniCPM-o team on the technical report; the GitHub LICENSE is the unfilled Apache template)",
      "grade": "C",
      "fit": "caveat",
      "why": "The template renders tool_call tags but the card never documents tool calling or names a parser; the Ollama build shows no tools badge.",
      "source": "https://huggingface.co/openbmb/MiniCPM-o-4_5/raw/main/chat_template.jinja",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/minicpm-o-4-5"
    },
    {
      "slug": "minicpm5-2b",
      "name": "MiniCPM5-2B",
      "company": "OpenBMB (named in the GitHub LICENSE's copyright line)",
      "grade": "C",
      "fit": "caveat",
      "why": "vLLM 0.29 and SGLang ship a minicpm5 parser for its XML calls and 1.6 GB fits any GPU, but thinking precedes each call, the Ollama page shows no tools badge, and discussions report a template crash on string arguments.",
      "source": "https://huggingface.co/openbmb/MiniCPM5-2B/raw/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/minicpm5-2b"
    },
    {
      "slug": "nemotron-3-5-bolt-e4b-4gb",
      "name": "NVIDIA Nemotron 3.5 Bolt E4B 4GB",
      "company": "NVIDIA Corporation",
      "grade": "C",
      "fit": "caveat",
      "why": "Built for voice assistants: fits 8 GB, thinking off, gemma4 parser, BFCL-v3 67.3 at Q4_K_M (NVIDIA's run). No account runs 30 commands, and Ollama's e4b tag advertises 128K against NVIDIA's 64K.",
      "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-4GB/raw/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-5-bolt-e4b-4gb"
    },
    {
      "slug": "nemotron-3-nano-4b",
      "name": "Nemotron 3 Nano 4B",
      "company": "NVIDIA Corporation",
      "grade": "C",
      "fit": "caveat",
      "why": "Fits an 8 GB GPU with thinking off, qwen3_coder parser in vLLM and an Ollama tools badge, but BFCL v3 is 61.1 on the card and no account measures 30 commands.",
      "source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16/raw/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-nano-4b"
    },
    {
      "slug": "ornith-1-5-9b",
      "name": "Ornith-1.5-9B",
      "company": "Ornith",
      "grade": "C",
      "fit": "caveat",
      "why": "Fits an 8 GB GPU and Ollama's 9b tag has a tools badge, but thinking is on by default and lengthy, so the 3-second median needs the undocumented template switch.",
      "source": "https://ollama.com/library/ornith-1.5/tags",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/ornith-1-5-9b"
    },
    {
      "slug": "qwen3-30b-a3b-instruct-2507",
      "name": "Qwen3-30B-A3B-Instruct-2507",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "C",
      "fit": "caveat",
      "why": "Non-thinking with tools in Ollama, but 19 GB needs Unsloth's expert offload to CPU to run on an 8 GB to 16 GB GPU, and no account measures latency or 30 commands.",
      "source": "https://ollama.com/library/qwen3/tags",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-30b-a3b-instruct-2507"
    },
    {
      "slug": "qwen3-4b-instruct-2507",
      "name": "Qwen3-4B-Instruct-2507",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "C",
      "fit": "caveat",
      "why": "Tools are in the template and the Ollama library declares tools, but no parser is named on the card and no account measures 30 commands on an 8 GB GPU.",
      "source": "https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507/raw/main/tokenizer_config.json",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-4b-instruct-2507"
    },
    {
      "slug": "qwen3-5-4b",
      "name": "Qwen3.5-4B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "C",
      "fit": "caveat",
      "why": "Parsers named and Ollama declares tools, thinking off per request; no account measures 30 commands on this size, and acc-28 shows an Ollama regression printing calls as text on the family.",
      "source": "https://huggingface.co/Qwen/Qwen3.5-4B",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-5-4b"
    },
    {
      "slug": "qwen3-5-9b",
      "name": "Qwen3.5-9B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "C",
      "fit": "caveat",
      "why": "acc-24 ran qwen3.5:9b in Home Assistant on an 8 GB GPU at 2.9 s median, but gives no success rate for it, and acc-26 found routing slow; parsers are named on the card.",
      "source": "https://huggingface.co/Qwen/Qwen3.5-9B",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-5-9b"
    },
    {
      "slug": "qwen3-coder-30b-a3b-instruct",
      "name": "Qwen3-Coder-30B-A3B-Instruct",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "C",
      "fit": "caveat",
      "why": "No thinking and a named parser, but Q4_K_M is 18.6 GB, so an 8 to 16 GB GPU offloads experts; no voice-pipeline account exists.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-coder-30b-a3b-instruct"
    },
    {
      "slug": "qwen3-vl-2b-instruct",
      "name": "Qwen3-VL-2B-Instruct",
      "company": "Alibaba Cloud (per the first batch's RDAP check of qwen.ai; this model's files name no copyright holder)",
      "grade": "C",
      "fit": "caveat",
      "why": "Tools in the template and Ollama declares tools, no thinking to block, 1.1 GB at Q4; the card names no parser and no account measures 30 commands.",
      "source": "https://huggingface.co/Qwen/Qwen3-VL-2B-Instruct/raw/main/chat_template.json",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-vl-2b-instruct"
    },
    {
      "slug": "qwen3-vl-4b-instruct",
      "name": "Qwen3-VL-4B-Instruct",
      "company": "Alibaba Cloud (per the first batch's RDAP check of qwen.ai; this model's files name no copyright holder)",
      "grade": "D",
      "fit": "caveat",
      "why": "Tools in the template and Ollama declares tools, no thinking to block, 2.5 GB at Q4; the card names no parser and no account measures 30 commands.",
      "source": "https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct/raw/main/chat_template.json",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-vl-4b-instruct"
    },
    {
      "slug": "qwen3-vl-8b-instruct",
      "name": "Qwen3-VL-8B-Instruct",
      "company": "Alibaba Cloud (per the pilot's RDAP check of qwen.ai; this model's files name no copyright holder)",
      "grade": "C",
      "fit": "caveat",
      "why": "Tools in the template and Ollama declares tools, no thinking to block, 5.0 GB at Q4; the card names no parser and no account measures 30 commands.",
      "source": "https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/raw/main/chat_template.json",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-vl-8b-instruct"
    },
    {
      "slug": "smollm3-3b",
      "name": "SmolLM3-3B",
      "company": "Hugging Face, Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "Card names vLLM's hermes parser and thinking off, 1.9 GB fits any GPU; no account measures 30 commands, and the template renders tool results as user turns.",
      "source": "https://huggingface.co/HuggingFaceTB/SmolLM3-3B",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/smollm3-3b"
    },
    {
      "slug": "spark-x2-6-4b",
      "name": "Spark-X2.6-4B",
      "company": "XHToken (the licence's copyright line); the Hugging Face organisation's name is SparkLLM",
      "grade": "C",
      "fit": "caveat",
      "why": "Aimed at agents with a named SGLang parser and thinking off per request on a 3.5 GB build, but the Ollama entry shows no tools badge and no account measures 30 commands.",
      "source": "https://huggingface.co/XHToken/Spark-X2.6-4B",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/spark-x2-6-4b"
    },
    {
      "slug": "apertus-v1-5-8b",
      "name": "Apertus v1.5 8B",
      "company": "ETH Zurich and EPFL (the Swiss National AI Institute, named in the Acceptable Use Policy)",
      "grade": "C",
      "fit": "no",
      "why": "Nothing runs it on an Ollama or llama.cpp server: the architecture is not in llama.cpp and Ollama has no library entry.",
      "source": "https://raw.githubusercontent.com/ggml-org/llama.cpp/master/src/llama-arch.cpp",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/apertus-v1-5-8b"
    },
    {
      "slug": "apodex-1-1-mini",
      "name": "Apodex-1.1-mini",
      "company": "Apodex (the site's copyright line reads 2026 Apodex)",
      "grade": "C",
      "fit": "no",
      "why": "The smallest publisher build is 21.7 GB, beyond an 8 to 16 GB GPU, and there is no Ollama tag.",
      "source": "https://huggingface.co/api/models/apodex/Apodex-1.1-mini-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/apodex-1-1-mini"
    },
    {
      "slug": "bigbang-v1",
      "name": "BigBang-v1",
      "company": "The Endless Frontier (endless-frontier)",
      "grade": "C",
      "fit": "no",
      "why": "The smallest build anywhere is 10 GB at IQ2_XXS from a third party and 21.9 GB at Q4; no Ollama entry and the card targets 8-GPU serving.",
      "source": "https://huggingface.co/api/models/bartowski/endless-frontier_BigBang-v1-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/bigbang-v1"
    },
    {
      "slug": "carbon-x",
      "name": "Carbon-X",
      "company": "Hugging Face (the HuggingFaceBio organisation, named Hugging Face Biology Research on the Hub)",
      "grade": "C",
      "fit": "no",
      "why": "No tool-calling training documented; the template inherits Qwen3's tool format but the card never uses it.",
      "source": "https://huggingface.co/HuggingFaceBio/Carbon-X",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/carbon-x"
    },
    {
      "slug": "cosmos-reason2-2b",
      "name": "Cosmos-Reason2-2B",
      "company": "NVIDIA Corporation",
      "grade": "C",
      "fit": "no",
      "why": "No tool-calling documentation; the inherited Qwen3-VL template carries tools but NVIDIA never exercises it.",
      "source": "https://huggingface.co/api/models/nvidia/Cosmos-Reason2-2B?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/cosmos-reason2-2b"
    },
    {
      "slug": "darwin-36b-opus",
      "name": "Darwin-36B-Opus",
      "company": "VIDRAFT Inc. (named on vidraft.net); the card's citation names FINAL-Bench and VIDRAFT_LAB",
      "grade": "C",
      "fit": "no",
      "why": "The Q4_K_M build is 21.2 GB; only the 1 to 3 bit POCKET files fit 16 GB, there is no Ollama entry and tool calling is undocumented.",
      "source": "https://huggingface.co/api/models/FINAL-Bench/POCKET-35B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/darwin-36b-opus"
    },
    {
      "slug": "deepseek-v3-2",
      "name": "DeepSeek-V3.2",
      "company": "DeepSeek (the MIT licence names DeepSeek, and DeepSeek's platform terms name Hangzhou DeepSeek Artificial Intelligence Co., Ltd.)",
      "grade": "C",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU; Ollama retired its deepseek-v3.2 entry on 15 July 2026.",
      "source": "https://huggingface.co/api/models/unsloth/DeepSeek-V3.2-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/deepseek-v3-2"
    },
    {
      "slug": "deepseek-v4-1-flash",
      "name": "DeepSeek V4.1 Flash",
      "company": "DeepSeek (the MIT licence names DeepSeek, and DeepSeek's platform terms name Hangzhou DeepSeek Artificial Intelligence Co., Ltd.)",
      "grade": "C",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU; Ollama's entry is cloud-only.",
      "source": "https://huggingface.co/api/models/unsloth/DeepSeek-V4.1-Flash-1017-GGUF?blobs=true",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/deepseek-v4-1-flash"
    },
    {
      "slug": "deepseek-v4-flash-0731",
      "name": "DeepSeek-V4-Flash-0731",
      "company": "DeepSeek (the MIT licence names DeepSeek, and DeepSeek's platform terms name Hangzhou DeepSeek Artificial Intelligence Co., Ltd.)",
      "grade": "C",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU; Ollama retired deepseek-v4-flash on 27 August 2026.",
      "source": "https://unsloth.ai/docs/models/deepseek-v4",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/deepseek-v4-flash-0731"
    },
    {
      "slug": "diffusiongemma-26b-a4b-it",
      "name": "DiffusionGemma 26B A4B IT",
      "company": "Google LLC (RDAP registrant of gemma.ai; the card names Google DeepMind as author)",
      "grade": "D",
      "fit": "no",
      "why": "Nothing serves it on Ollama or released llama.cpp; the 16.8 GB third-party Q4 needs an unmerged branch.",
      "source": "https://huggingface.co/unsloth/diffusiongemma-26B-A4B-it-GGUF",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/diffusiongemma-26b-a4b-it"
    },
    {
      "slug": "gemma-4-31b-it",
      "name": "Gemma 4 31B IT",
      "company": "Google LLC (the registrant of gemma.ai; the card names Google DeepMind as author)",
      "grade": "BB",
      "fit": "no",
      "why": "17.6 GB at Q4_0 exceeds an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/google/gemma-4-31B-it-qat-q4_0-gguf?blobs=true",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gemma-4-31b-it"
    },
    {
      "slug": "glm-4-5-air",
      "name": "GLM-4.5-Air",
      "company": "JINGSHENG HENGXING TECHNOLOGY PTE.LTD (operator of Z.ai per its terms of use, and registrant of z.ai, per the first batch's check on 21 October 2026)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 to 16 GB GPU and Ollama has no entry.",
      "source": "https://huggingface.co/api/models/unsloth/GLM-4.5-Air-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/glm-4-5-air"
    },
    {
      "slug": "glm-5-3",
      "name": "GLM-5.3",
      "company": "JINGSHENG HENGXING TECHNOLOGY PTE.LTD (operator of Z.ai per its terms of use, and registrant of z.ai). The licence names Z.AI",
      "grade": "B",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU; no Ollama entry.",
      "source": "https://huggingface.co/api/models/unsloth/GLM-5.3-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/glm-5-3"
    },
    {
      "slug": "glm-5-3-flash",
      "name": "GLM-5.3-Flash",
      "company": "Z.AI Co., Ltd (named in the licence's copyright line)",
      "grade": "B",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU; Ollama's entry is cloud-only.",
      "source": "https://huggingface.co/api/models/unsloth/GLM-5.3-Flash-GGUF?blobs=true",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/glm-5-3-flash"
    },
    {
      "slug": "gpt-oss-120b",
      "name": "gpt-oss-120b",
      "company": "OpenAI (named as publisher on the card; the openai-api listing's terms name OpenAI OpCo, LLC)",
      "grade": "BB",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/ggml-org/gpt-oss-120b-GGUF?blobs=true",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gpt-oss-120b"
    },
    {
      "slug": "hy3",
      "name": "Hy3",
      "company": "Tencent (named in the LICENSE file's copyright line)",
      "grade": "B",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU; no Ollama entry.",
      "source": "https://huggingface.co/api/models/AngelSlim/Hy3-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/hy3"
    },
    {
      "slug": "inkling",
      "name": "Inkling",
      "company": "Thinking Machines Lab, Inc. (named on its model card page and in its training-content summary)",
      "grade": "B",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU.",
      "source": "https://unsloth.ai/docs/models/inkling",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/inkling"
    },
    {
      "slug": "inkling-small",
      "name": "Inkling-Small",
      "company": "Thinking Machines Lab, Inc. (named on its model card page and in its training-content summary)",
      "grade": "B",
      "fit": "no",
      "why": "Cannot load on an 8 to 16 GB GPU; no Ollama entry.",
      "source": "https://huggingface.co/api/models/unsloth/Inkling-Small-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/inkling-small"
    },
    {
      "slug": "internvl3-5-gpt-oss-20b-a4b-preview",
      "name": "InternVL3.5-GPT-OSS-20B-A4B-Preview",
      "company": "Shanghai AI Laboratory (OpenGVLab), not named in the repository",
      "grade": "D",
      "fit": "no",
      "why": "No tool calling in the template or card.",
      "source": "https://huggingface.co/OpenGVLab/InternVL3_5-GPT-OSS-20B-A4B-Preview-HF/raw/main/chat_template.jinja",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/internvl3-5-gpt-oss-20b-a4b-preview"
    },
    {
      "slug": "kimi-k3",
      "name": "Kimi K3",
      "company": "Moonshot AI (the licence names Moonshot AI; the Kimi Platform terms name Moonshot AI PTE. LTD.)",
      "grade": "B",
      "fit": "no",
      "why": "Cannot load locally on a small GPU; Ollama's kimi-k3 is cloud-only.",
      "source": "https://ollama.com/library/kimi-k3/tags",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/kimi-k3"
    },
    {
      "slug": "kolibri-1",
      "name": "Kolibri-1",
      "company": "Aleph Alpha GmbH",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 to 16 GB GPU and Ollama has no entry.",
      "source": "https://huggingface.co/api/models/bartowski/Kolibri-1-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/kolibri-1"
    },
    {
      "slug": "kronos",
      "name": "Kronos",
      "company": "No company. Yu Shi publishes it from the GitHub account shiyu-coder and the Hugging Face user account NeoQuasar (named ShiYu), with co-authors at Tsinghua University per the paper",
      "grade": "C",
      "fit": "no",
      "why": "No tool calling; the model takes OHLCV bars and returns forecast bars.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/kronos"
    },
    {
      "slug": "laguna-m-1",
      "name": "Laguna M.1",
      "company": "Poolside, Inc. (named in Poolside's Terms of Use; the LICENSE.md copyright line reads Poolside)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/laguna-m-1"
    },
    {
      "slug": "laguna-s-2-1",
      "name": "Laguna S 2.1",
      "company": "Poolside, Inc. (named in Poolside's Terms of Use)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU.",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/laguna-s-2-1"
    },
    {
      "slug": "laguna-xs-2-1",
      "name": "Laguna XS 2.1",
      "company": "Poolside, Inc. (named in Poolside's Terms of Use)",
      "grade": "C",
      "fit": "no",
      "why": "The only quantised builds are 20 GB Q4_K_M and larger; nothing fits an 8 GB to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/poolside/Laguna-XS-2.1-GGUF/tree/main?recursive=true",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/laguna-xs-2-1"
    },
    {
      "slug": "lbs-27b",
      "name": "LBS-27B",
      "company": "inclusionAI, Ant Group (named in the site footer of inclusion-ai.org); the work is credited to authors from Wuhan University of Technology, CASIA, Zhejiang University and Ant Group",
      "grade": "C",
      "fit": "no",
      "why": "No tool calling; actions are plain-text crop instructions parsed by inclusionAI's Python.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/lbs-27b"
    },
    {
      "slug": "minimax-m2-5",
      "name": "MiniMax-M2.5",
      "company": "MiniMax AI (named in the licences; MiniMax's platform terms name Nanonoble Pte. Ltd., Singapore)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/minimax-m2-5"
    },
    {
      "slug": "minimax-m2-7",
      "name": "MiniMax-M2.7",
      "company": "MiniMax (named in the licence's copyright line; MiniMax's platform terms name Nanonoble Pte. Ltd., Singapore)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU, and every reply opens a think block.",
      "source": "https://huggingface.co/api/models/unsloth/MiniMax-M2.7-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/minimax-m2-7"
    },
    {
      "slug": "minimax-m3-1-flash",
      "name": "MiniMax-M3.1-Flash",
      "company": "MiniMax (named in the licence's copyright line; MiniMax's platform terms name Nanonoble Pte. Ltd., Singapore)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU, and thinking is adaptive with no off switch in the template.",
      "source": "https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash/raw/main/chat_template.jinja",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/minimax-m3-1-flash"
    },
    {
      "slug": "mistral-large-4",
      "name": "Mistral Large 4",
      "company": "Mistral AI (RCS Paris 952 418 325)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU, and reasoning_effort has no off value.",
      "source": "https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B/raw/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/mistral-large-4"
    },
    {
      "slug": "muse-glimmer-30b",
      "name": "Muse Glimmer 30B",
      "company": "Meta (the verified Hugging Face organisation is \"Meta Inc.\", the card names Meta Superintelligence Lab, and the licence carries no copyright line)",
      "grade": "BB",
      "fit": "no",
      "why": "Smallest build is 16.8 GB (18 GB in Ollama), over an 8 GB GPU, and thinking cannot be turned off.",
      "source": "https://huggingface.co/api/models/meta-models/Muse-Glimmer-30B-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/muse-glimmer-30b"
    },
    {
      "slug": "nanbeige-4-3-33b",
      "name": "Nanbeige4.3-33B",
      "company": "BOSS Zhipin (Kanzhun Limited), unconfirmed",
      "grade": "C",
      "fit": "no",
      "why": "Smallest GGUF is 10.1 GB at IQ2_XXS; nothing at a usable bit width fits an 8 GB GPU, and there is no Ollama library entry.",
      "source": "https://huggingface.co/api/models/bartowski/Nanbeige_Nanbeige4.3-33B-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nanbeige-4-3-33b"
    },
    {
      "slug": "nemotron-3-5-lightning-30b-a3b",
      "name": "Nemotron 3.5 Lightning 30B A3B",
      "company": "NVIDIA Corporation",
      "grade": "B",
      "fit": "no",
      "why": "Smallest build is 18.9 GB (Ollama 19 GB), over an 8 GB to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/ggml-org/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-5-lightning-30b-a3b"
    },
    {
      "slug": "nemotron-3-nano-30b-a3b",
      "name": "Nemotron 3 Nano 30B A3B",
      "company": "NVIDIA Corporation",
      "grade": "B",
      "fit": "no",
      "why": "Smallest Ollama build is 18 GB, over an 8 GB to 16 GB GPU.",
      "source": "https://ollama.com/library/nemotron-3-nano",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-nano-30b-a3b"
    },
    {
      "slug": "nemotron-3-nano-omni-30b-a3b-reasoning",
      "name": "Nemotron 3 Nano Omni 30B A3B Reasoning",
      "company": "NVIDIA Corporation",
      "grade": "C",
      "fit": "no",
      "why": "Smallest build is 18.5 GB (Ollama 21 GB), over an 8 GB to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/unsloth/NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-nano-omni-30b-a3b-reasoning"
    },
    {
      "slug": "nemotron-3-super-120b-a12b",
      "name": "Nemotron 3 Super 120B A12B",
      "company": "NVIDIA Corporation",
      "grade": "B",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU.",
      "source": "https://unsloth.ai/docs/models/nemotron-3/nemotron-3-super",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-super-120b-a12b"
    },
    {
      "slug": "nemotron-3-ultra-550b-a55b",
      "name": "Nemotron 3 Ultra 550B A55B",
      "company": "NVIDIA Corporation",
      "grade": "B",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU; Ollama lists only a cloud tag.",
      "source": "https://huggingface.co/api/models/unsloth/NVIDIA-Nemotron-3-Ultra-550B-A55B-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-ultra-550b-a55b"
    },
    {
      "slug": "occamy-1-0",
      "name": "occamy-1.0",
      "company": "Accio (Accio-Lab)",
      "grade": "C",
      "fit": "no",
      "why": "Smallest GGUF is 11.7 GB, over an 8 GB GPU, and there is no Ollama library entry.",
      "source": "https://huggingface.co/api/models/Accio-Lab/occamy-1.0-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/occamy-1-0"
    },
    {
      "slug": "olmo-3-7b-think",
      "name": "Olmo 3 7B Think",
      "company": "The Allen Institute for Artificial Intelligence (Ai2), a 501(c)(3) nonprofit, per its Responsible Use Guidelines",
      "grade": "D",
      "fit": "no",
      "why": "Not wired for OpenAI-style tools (custom functions field, no parser) and thinking cannot be turned off.",
      "source": "https://huggingface.co/allenai/Olmo-3-7B-Think/raw/main/chat_template.jinja",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/olmo-3-7b-think"
    },
    {
      "slug": "ornith-1-5-35b-a3b",
      "name": "Ornith-1.5-35B-A3B",
      "company": "Ornith",
      "grade": "C",
      "fit": "no",
      "why": "Smallest build is Q4_K_M at 21.7 GB; nothing fits an 8 GB to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/ornith-ai/Ornith-1.5-35B-A3B-GGUF/tree/main?recursive=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/ornith-1-5-35b-a3b"
    },
    {
      "slug": "ornith-1-5-397b",
      "name": "Ornith-1.5-397B",
      "company": "Ornith",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 GB to 16 GB GPU.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/ornith-1-5-397b"
    },
    {
      "slug": "qwen3-5-122b-a10b",
      "name": "Qwen3.5-122B-A10B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "B",
      "fit": "no",
      "why": "No build fits an 8 to 16 GB GPU; the smallest Unsloth quant is 34 GB.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3.5-122B-A10B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-5-122b-a10b"
    },
    {
      "slug": "qwen3-5-397b-a17b",
      "name": "Qwen3.5-397B-A17B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "B",
      "fit": "no",
      "why": "No build fits an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3.5-397B-A17B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-5-397b-a17b"
    },
    {
      "slug": "qwen3-6-27b",
      "name": "Qwen3.6-27B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "B",
      "fit": "no",
      "why": "IQ4_XS is 15.4 GB before 4 GB of KV at 64K, so it does not fit an 8 to 16 GB GPU, and acc-11 saw it describe actions instead of calling tools on Ollama.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3.6-27B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-6-27b"
    },
    {
      "slug": "qwen3-6-35b-a3b",
      "name": "Qwen3.6-35B-A3B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "B",
      "fit": "no",
      "why": "UD-Q4_K_XL is 22 GB, past an 8 to 16 GB GPU, and acc-23 saw the family's 35B fail tool calling on Ollama before it worked on llama.cpp.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3.6-35B-A3B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-6-35b-a3b"
    },
    {
      "slug": "qwen3-8-27b",
      "name": "Qwen3.8-27B",
      "company": "Alibaba Cloud (named in the licence's copyright line)",
      "grade": "B",
      "fit": "no",
      "why": "UD-IQ4_XS is 14.4 GB before 4 GB of KV at 64K; too large for an 8 GB GPU and no account measures it in a voice pipeline.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3.8-27B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-8-27b"
    },
    {
      "slug": "qwen3-8-flash-next",
      "name": "Qwen3.8-Flash-Next",
      "company": "Alibaba Cloud (Singapore) Private Limited, the registrant of qwen.ai. The licence names Qwen as copyright holder",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3.8-Flash-Next-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-8-flash-next"
    },
    {
      "slug": "qwen3-coder-next",
      "name": "Qwen3-Coder-Next",
      "company": "Alibaba Cloud (the Qwen team's employer; no LICENSE file names a copyright holder for this model)",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 to 16 GB GPU; Unsloth's smallest is 19 GB at 1 bit.",
      "source": "https://huggingface.co/api/models/unsloth/Qwen3-Coder-Next-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-coder-next"
    },
    {
      "slug": "qwen3-omni-30b-a3b-instruct",
      "name": "Qwen3-Omni-30B-A3B-Instruct",
      "company": "Alibaba Cloud (carried from the 20 October 2026 pilot; this model's files name no copyright holder)",
      "grade": "D",
      "fit": "no",
      "why": "Tool calls are JSON in the template with no parser named, and the smallest build is 18.6 GB plus a 1.3 GB projector, past an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/Qwen/Qwen3-Omni-30B-A3B-Instruct/raw/main/chat_template.json",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-omni-30b-a3b-instruct"
    },
    {
      "slug": "qwen3-vl-30b-a3b-instruct",
      "name": "Qwen3-VL-30B-A3B-Instruct",
      "company": "Alibaba Cloud (per the pilot's RDAP check of qwen.ai; this model's files name no copyright holder)",
      "grade": "C",
      "fit": "no",
      "why": "18.6 GB at Q4 plus projector does not fit an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/Qwen/Qwen3-VL-30B-A3B-Instruct-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-vl-30b-a3b-instruct"
    },
    {
      "slug": "qwen3-vl-32b-instruct",
      "name": "Qwen3-VL-32B-Instruct",
      "company": "Alibaba Cloud (per the first batch's RDAP check of qwen.ai; this model's files name no copyright holder)",
      "grade": "C",
      "fit": "no",
      "why": "19.8 GB at Q4 does not fit an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/Qwen/Qwen3-VL-32B-Instruct-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwen3-vl-32b-instruct"
    },
    {
      "slug": "qwopus3-8-27b-flash-v2",
      "name": "Qwopus3.8-27B-Flash-V2",
      "company": "Jackrong, an individual Hugging Face account",
      "grade": "C",
      "fit": "no",
      "why": "IQ4_XS is 15.4 GB before KV and the recommended mode is thinking at xhigh; no fit for an 8 GB GPU voice pipeline.",
      "source": "https://huggingface.co/api/models/Jackrong/Qwopus3.8-27B-Flash-V2-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qwopus3-8-27b-flash-v2"
    },
    {
      "slug": "recursive-flint",
      "name": "Recursive Flint",
      "company": "Recursive Superintelligence, Inc. (the Hugging Face organisation's full name)",
      "grade": "C",
      "fit": "no",
      "why": "Thinking cannot be switched off, which the pipeline cannot tolerate, and 23 GB passes an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/RecursiveSI/flint/raw/main/sglang/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/recursive-flint"
    },
    {
      "slug": "ripple-65b-preview",
      "name": "Ripple-65B Preview",
      "company": "Ricursive Intelligence, Inc.",
      "grade": "D",
      "fit": "no",
      "why": "Thinking cannot be turned off and nothing fits an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/ricursiveAI/Ripple-65B-Preview/raw/main/inference/server/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/ripple-65b-preview"
    },
    {
      "slug": "swift-1-5-qwen3-8-27b",
      "name": "Swift 1.5 Qwen3.8-27B",
      "company": "UkisAI",
      "grade": "C",
      "fit": "no",
      "why": "IQ4_XS is 15.5 GB before KV; too large for an 8 GB GPU.",
      "source": "https://huggingface.co/api/models/ukisai/Swift-1.5-Qwen3.8-27B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/swift-1-5-qwen3-8-27b"
    },
    {
      "slug": "swift-1-5-qwen3-8-flash-next",
      "name": "Swift 1.5 Qwen3.8-Flash-Next",
      "company": "UkisAI",
      "grade": "C",
      "fit": "no",
      "why": "No build fits an 8 to 16 GB GPU.",
      "source": "https://huggingface.co/api/models/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/swift-1-5-qwen3-8-flash-next"
    },
    {
      "slug": "thinkingcap-qwen3-8-27b",
      "name": "ThinkingCap (Qwen3.8 27B)",
      "company": "BottleCap AI (named in the licence's copyright line)",
      "grade": "C",
      "fit": "no",
      "why": "IQ4_XS is 15.5 GB before KV; too large for an 8 GB GPU.",
      "source": "https://huggingface.co/api/models/bottlecapai/ThinkingCap-Qwen3.8-27B-GGUF?blobs=true",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/thinkingcap-qwen3-8-27b"
    },
    {
      "slug": "unlimited-ocr-2",
      "name": "Baidu Unlimited-OCR-2",
      "company": "Baidu, Inc.",
      "grade": "C",
      "fit": "no",
      "why": "No tool calling and no chat template.",
      "source": "https://huggingface.co/baidu/Unlimited-OCR-2/raw/main/tokenizer_config.json",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/unlimited-ocr-2"
    }
  ],
  "marks": "yes: the docs show it does the job as the success test says. caveat: it does the job with a limit an agent should know. no: it doesn't do this job. unchecked: we couldn't read it.",
  "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": "2026-11-16",
    "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"
  },
  "useCase": {
    "accounts": 7,
    "after": [
      "home-automation.execute",
      "text-to-speech.speak"
    ],
    "area": "models",
    "asks": [
      "run Home Assistant Assist on a local model",
      "a voice assistant that turns the lights off",
      "route a spoken question to the right tool",
      "an 8 GB GPU model that never misses a tool call"
    ],
    "before": [
      "speech-to-text.transcribe"
    ],
    "blockers": [
      "Ollama template lacks tools, so the server answers 'does not support tools'",
      "a 3-bit quant gets stuck in hidden reasoning and returns nothing",
      "runtime regression prints the tool call as text"
    ],
    "capabilities": [
      "model.open-weights",
      "model.tool-calling",
      "model.thinking-off"
    ],
    "category": "open-weights",
    "categoryName": "Open-weight models",
    "checked": "2026-11-12",
    "companies": {
      "fit": 2,
      "fitWithLimit": 16,
      "tagged": 90
    },
    "competition": "few",
    "fits": [
      "granite-4-2-3b",
      "qwen3-5-0-8b",
      "qwen3-5-2b"
    ],
    "fitsWithLimit": [
      "agents-a1-4b",
      "gemma-4-12b-it",
      "gemma-4-26b-a4b-it",
      "gemma-4-e2b-it",
      "gemma-4-e4b-it",
      "glm-4-7-flash",
      "gpt-oss-20b",
      "lfm2-5-2-6b",
      "ling-3-0-tiny",
      "mimo-v2-6-distill-qwen-9b",
      "minicpm-o-4-5",
      "minicpm5-2b",
      "nemotron-3-5-bolt-e4b-4gb",
      "nemotron-3-nano-4b",
      "ornith-1-5-9b",
      "qwen3-30b-a3b-instruct-2507",
      "qwen3-4b-instruct-2507",
      "qwen3-5-4b",
      "qwen3-5-9b",
      "qwen3-coder-30b-a3b-instruct",
      "qwen3-vl-2b-instruct",
      "qwen3-vl-4b-instruct",
      "qwen3-vl-8b-instruct",
      "smollm3-3b",
      "spark-x2-6-4b"
    ],
    "id": "open-weights.tool-calling-assistant",
    "listings": {
      "caveat": 25,
      "no": 58,
      "notRead": 28,
      "tagged": 161,
      "unchecked": 0,
      "yes": 3
    },
    "name": "A local assistant that calls tools in under 3 seconds",
    "status": "draft",
    "success": "The model returns tool calls with JSON arguments, running on an 8 GB GPU. Measured: at least 95 percent of 30 test commands produce the right tool call with valid JSON arguments, median latency under 3 seconds on an 8 GB GPU, and zero empty replies.",
    "url": "https://www.anchorterminal.com/api/v1/use-cases/open-weights.tool-calling-assistant.json",
    "weight": 10,
    "weightBasis": "7 of 51 accounts, all from the Home Assistant forum, personal blogs and Ollama issues"
  }
}
