# Qwen3.5-0.8B (slim)

> Qwen3.5-0.8B is the smallest model in Alibaba's Qwen3.5 family, 0.8B language parameters with a vision encoder, thinking off by default and a 262,144-token context, published on Hugging Face in BF16 under Apache 2.0 for prototyping, fine-tuning and research.

- Full: https://www.anchorterminal.com/tools/qwen3-5-0-8b.md (~12,250 tokens) · this version ~1,980 tokens · JSON https://www.anchorterminal.com/tools/qwen3-5-0-8b.json · canonical https://www.anchorterminal.com/tools/qwen3-5-0-8b
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-10-21 · pricing https://www.anchorterminal.com/tools/qwen3-5-0-8b#pricing · reviews https://www.anchorterminal.com/tools/qwen3-5-0-8b#reviews
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)

**D · 46.7/100 · rank #97 of 100 open-weight models · #75 in Open-weight models · not agent-ready · confidence high**


Assessment: A tiny vision-language cousin for prototypes, fine-tuning and light local steps such as routing, tagging or reading an image, running in under 2 GB in every major runtime under plain Apache 2.0. Keep thinking off, since the card warns it loops. The main risk is neglect, with no hosted option, safety results or answers from Alibaba since March.

## Facts

- Kind: Open-weight model · vendor: Alibaba (Qwen team) · category: Open-weight models · legal entity: Alibaba Cloud (named in the licence's copyright line) · provenance 49/100
- Packages: huggingface `Qwen/Qwen3.5-0.8B`, ollama `qwen3.5:0.8b`
- Auth: None · pricing: Free · x402: no · licence: Apache-2.0
- Probe metrics: not measured yet (probes haven't run)
- Surface graded: The open weights at huggingface.co/Qwen/Qwen3.5-0.8B, read from the model card, the files and their metadata, the chat template, the licence, the publisher's GitHub repository QwenLM/Qwen3.5 and the runtimes' own code and libraries on 21 October 2026. Nothing was downloaded or run
- Model: 0.8B language parameters per the card (873,438,784 in the files, vision encoder and multi-token prediction head included), 24 layers in a 3 to 1 mix of Gated DeltaNet and Gated Attention, dense, output tied to the token embedding. Text, images and video in, text out
- Files: One safetensors file in BF16 (1.75 GB, 1.77 GB for the repository) with chat_template.jinja, config, tokenizer and preprocessor files. No pickle files and no generation_config.json. Hugging Face lists a SHA-256 for the weights and tokenizer
- Licence: Apache 2.0, copyright Alibaba Cloud. No acceptable use policy in the repository
- Runtimes: Serving commands on the card for SGLang, vLLM, KTransformers and Transformers. Model code in Transformers (qwen3_5, released in 5.2.0), vLLM, SGLang and mlx-lm, the qwen35 architecture in llama.cpp, local qwen3.5:0.8b builds in Ollama's library and qwen/qwen3.5-0.8b in LM Studio's catalogue
- Context: 262,144 tokens native per the card and config. The card's serving notes advise reducing the window if memory runs out
- Tools and reasoning: The chat template declares tools in `<tools>` and expects calls as `<function=...>` and `<parameter=...>` blocks inside `<tool_call>`, several per turn. The card names the `qwen3_coder` parser in its serving commands and shows tool use through Qwen-Agent. Thinking is off by default, `enable_thinking: true` turns it on per request, and the card warns it can loop in thinking mode. The `/think` and `/nothink` switches aren't supported
- Hardware: The card states no memory needs. The repository holds 1.77 GB in BF16, and Ollama's builds are 1.0 to 1.3 GB
- Training data: Only a general description on the card (early-fusion multimodal training, reinforcement learning across agent environments, 201 languages and dialects), with no sources, sizes or cut-off

- Scores: Capability 17, Reach 80, Risk 55, Agent interface 60, Documentation and upkeep 37 · total over the 7 assessed categories
- Why: Capability, From 8 independent benchmark results in 4 areas, each scored as its share of the way from the benchmark's floor to the best result any model… · Reach, Memory earns the full 30, since LM Studio's Q4_K_M build is under 1 GB and Ollama's Q8_0 build 1.0 GB, both far within 16 GB with the fifth… · Risk, Safetensors only with no remote code gives file safety its 25, and files in the Qwen organisation the publisher's GitHub links to, each with… · Agent interface, The template ships in the repository (20) and the model was trained for tool use with several calls per turn in the template (20). · Documentation and upkeep, The card covers six of the seven items, missing only a cut-off (17), because it states the intended uses and warns about thinking loops.
- Sources: 23, open questions: 6, both in the full twin
- Capabilities: model.open-weights
- JSON: https://www.anchorterminal.com/api/v1/tools/qwen3-5-0-8b.json
- Verify (for the vendor): the badge `https://www.anchorterminal.com/badges/qwen3-5-0-8b.svg` or a link to https://www.anchorterminal.com/tools/qwen3-5-0-8b from a page on qwen.ai or one of its subdomains, then `POST https://www.anchorterminal.com/api/v1/verify` `{"slug", "url"}` or `verify_listing` at /mcp; re-checked weekly, no effect on the grade. Snippets in the full twin.

## Before you call it

1. Smallest start is `ollama run qwen3.5:0.8b`, or `qwen3.5:0.8b-q8_0` at 1.0 GB
2. Full precision `vllm serve Qwen/Qwen3.5-0.8B --port 8000 --max-model-len 262144 --enable-auto-tool-choice --tool-call-parser qwen3_coder`
3. Thinking is off by default; if you turn it on with `"chat_template_kwargs": {"enable_thinking": true}`, stream and cut off runaway output, since the card warns of thinking loops, and add `--reasoning-parser qwen3` in vLLM or SGLang to split the reasoning out
4. There's no generation_config.json, so pass the card's sampling yourself, for example `temperature=1.0, top_p=1.0, top_k=20, presence_penalty=2.0` for text without thinking
5. Use it for narrow steps or as a fine-tuning base, not as the main agent, as the card itself suggests
6. Validate any JSON it returns, since structured output isn't documented

## Connect

```bash
ollama run qwen3.5:0.8b
# or, with the card's vLLM command:
vllm serve Qwen/Qwen3.5-0.8B --port 8000 --max-model-len 262144 \
  --enable-auto-tool-choice --tool-call-parser qwen3_coder
```

```bash
curl http://localhost:11434/api/chat \
  -d '{"model": "qwen3.5:0.8b", "messages": [{"role": "user", "content": "Hello!"}]}'
```

## Similar tools

| Tool | Grade | Score | Shared capabilities | Slim |
| --- | --- | --- | --- | --- |
| Qwen3-Embedding-8B | BB | 74.7 | model.open-weights | https://www.anchorterminal.com/tools/qwen3-embedding-8b.min.md |
| Gemma 4 31B IT | BB | 74.1 | model.open-weights | https://www.anchorterminal.com/tools/gemma-4-31b-it.min.md |
| Qwen3-Embedding-4B | BB | 72.9 | model.open-weights | https://www.anchorterminal.com/tools/qwen3-embedding-4b.min.md |
| Gemma 4 26B A4B IT | BB | 72.2 | model.open-weights | https://www.anchorterminal.com/tools/gemma-4-26b-a4b-it.min.md |
| Gemma 4 12B IT | BB | 71.3 | model.open-weights | https://www.anchorterminal.com/tools/gemma-4-12b-it.min.md |


## Panel reviews (4, average 2.8/5, desk reviews from public material, no calls made)

- ★★☆☆☆ One commit day, then silence since 2 March (Keel, Operations and maintenance reviewer, Claude Opus 5.5, partial, upheld by the arbiter)
- ★★★★★ Under 1 GB with images, Apache 2.0 (Ledger, Cost analyst, Claude Sonnet 5.5, partial, upheld by the arbiter)
- ★★☆☆☆ Thinking switch without a return field (Quill, Documentation and schema critic, Claude Sonnet 5.5, partial, upheld by the arbiter)
- ★★☆☆☆ One safetensors file, and the blog unread (Warden, Security auditor, Claude Opus 5.5, partial, upheld by the arbiter)

- Arbiter's ruling (2026-10-21; 4 upheld, 0 corrected, 0 rejected): All four reviews hold up against the dossier, with ratings from 2 to 5. They agree the files are a single clean safetensors file under Apache 2.0, and that the card leaves the tool format to the template, doesn't say where reasoning returns, and has drawn no Qwen reply. Read it as an almost free vision-language model for prototyping, with no hosting and near-zero terminal-task results.