Jackrong

/

Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled

like 2.74k

ollama run hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF:Q4_K_M

ollama run hf.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF:Q4_K_M

Error: 500 Internal Server Error: unable to load model :(

Ollama hasn't caught up with supporting Qwen 3.5 yet; you can use LM Studio instead.

Ollama hasn't caught up with supporting Qwen 3.5 yet; you can use LM Studio instead.

what? Qwen 3.5 is supported from the official repo https://ollama.com/library/qwen3.5

I believe this has to do with the model "architecture" being set to qwen35 for the ones in ollama's repos, vs. this being Qwen3_5ForConditionalGeneration (or qwen3_5 - I'm not sure which one of these is actually used, but your server logs will say.)

I expect Ollama will update soon to allow these, but, in the meantime, you can use llama.cpp directly if you'd like - which is what ollama does, more or less. (It's based on llama.cpp)

可以用ollama0.18.2版本来运行，已成功在vscode里用claude code for vscode调用。我用的是wsl+ubuntu，先下载Qwen3.5-9B.Q8_0.gguf到本地，。先创建Modelfile.qwen35-local-tools文件 ，写入内容如下：
FROM Qwen3.5-9B.Q8_0.gguf:latest #根据你的模型来更改名称
TEMPLATE {{ .Prompt }}
RENDERER qwen3.5
PARSER qwen3.5
PARAMETER num_ctx 32768
然后运行ollama导入Qwen3.5-9B.Q8_0.gguf 。然后查看模型名称ollama list .
再在claude code for vscode的settings.json里填入：
{
"name": "ANTHROPIC_BASE_URL",
"value": "http://127.0.0.1:11434"
},
{
"name": "ANTHROPIC_AUTH_TOKEN",
"value": "ollama"
},
{
"name": "ANTHROPIC_API_KEY",
"value": ""
},
{
"name": "ANTHROPIC_MODEL",
"value": "qwen35-local-tools:latest"
},
{
"name": "ANTHROPIC_SMALL_FAST_MODEL",
"value": "qwen35-local-tools:latest"
},
{
"name": "ANTHROPIC_DEFAULT_SONNET_MODEL",
"value": "qwen35-local-tools:latest"
},
{
"name": "ANTHROPIC_DEFAULT_OPUS_MODEL",
"value": "qwen35-local-tools:latest"
},
{
"name": "ANTHROPIC_CUSTOM_MODEL_OPTION",
"value": "qwen35-local-tools:latest"
},
{
"name": "ANTHROPIC_CUSTOM_MODEL_OPTION_NAME",
"value": "qwen35-local-tools(local)"
} 注意模型名与前面一致。重启claude code for vscode 就可以用了。

Ollama hasn't caught up with supporting Qwen 3.5 yet; you can use LM Studio instead.

what? Qwen 3.5 is supported from the official repo https://ollama.com/library/qwen3.5

You need to look at this in the context of the question; what I mean is Qwen3.5 in GGUF format

Its recommended not to use Ollama now. You should switch to LM Studio CLI or just raw LLamaCPP.