- Size
- 2.3B parameterscounted from the safetensors header on the Hub
- Licence
- custom, read it, Apache License 2.0Our reading, check the licence: A custom licence we have not read for this model: other makers use the same name for different texts. Read the model's own licence file before you use it.Licence text
- On the Hub since
- 28 Feb 2026the date the repo was created
- Context length
- 256k tokensfrom the model's config
- Fine-tune
- of Qwen/Qwen3.5-2B-BaseQwen3_5ForConditionalGeneration
- On Hugging Face
- Qwen/Qwen3.5-2B5.5M downloads in the last 30 days, 492 likes
Scores
- effort none0.0%via Epoch AI
Epoch's name: qwen3.5-2b. Best closed model on this set: gemini-4-carbon_high, 36.0% (Google DeepMind). All 271 results
Claimed by the maker Claimed
The model card has no claimed scores in a machine-readable form (the card's model-index), so none are shown. The maker may give scores in the card text: read the card.
Other views, which we link to and do not copy: look this model up on Artificial Analysis and the Arena leaderboard.
Independent scores from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Each row links its own source. What the badges mean.
Run it yourself
Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 0.09 GB, worked out from the model's config (Qwen/Qwen3.5-2B), at 16-bit precision. Choose a machine above to see a verdict for each format. Method
GGUF for llama.cpp, Ollama, LM Studio and others
unsloth/Qwen3.5-2B-GGUF (converted by unsloth), 534k downloads. Sizes are the real files on the Hub, all parts added up. For image input you also need the projector file mmproj-F16.gguf (0.62 GB), not counted below.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| UD-IQ3_XXS | 0.87 GBfile size | 1.5 GB0.87 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:UD-IQ3_XXSOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:UD-IQ3_XXSDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-UD-IQ3_XXS.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q3_K_S | 0.96 GBfile size | 1.6 GB0.96 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q3_K_SOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q3_K_SDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q3_K_S.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q3_K_M | 1.0 GBfile size | 1.6 GB1.0 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q3_K_MOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q3_K_MDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q3_K_M.gguf" --local-dir Qwen3.5-2B-GGUF |
| UD-Q3_K_XL | 1.1 GBfile size | 1.7 GB1.1 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:UD-Q3_K_XLOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:UD-Q3_K_XLDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-UD-Q3_K_XL.gguf" --local-dir Qwen3.5-2B-GGUF |
| IQ4_XS | 1.1 GBfile size | 1.7 GB1.1 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:IQ4_XSOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:IQ4_XSDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-IQ4_XS.gguf" --local-dir Qwen3.5-2B-GGUF |
| IQ4_NL | 1.1 GBfile size | 1.7 GB1.1 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:IQ4_NLOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:IQ4_NLDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-IQ4_NL.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q4_0 | 1.1 GBfile size | 1.7 GB1.1 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q4_0Ollama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q4_0Download hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q4_0.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q4_K_S | 1.1 GBfile size | 1.7 GB1.1 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q4_K_SOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q4_K_SDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q4_K_S.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q4_K_M | 1.2 GBfile size | 1.8 GB1.2 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q4_K_MOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q4_K_MDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q4_K_M.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q4_1 | 1.2 GBfile size | 1.8 GB1.2 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q4_1Ollama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q4_1Download hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q4_1.gguf" --local-dir Qwen3.5-2B-GGUF |
| UD-Q4_K_XL | 1.2 GBfile size | 1.8 GB1.2 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:UD-Q4_K_XLOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:UD-Q4_K_XLDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-UD-Q4_K_XL.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q5_K_S | 1.3 GBfile size | 1.9 GB1.3 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q5_K_SOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q5_K_SDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q5_K_S.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q5_K_M | 1.3 GBfile size | 1.9 GB1.3 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q5_K_MOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q5_K_MDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q5_K_M.gguf" --local-dir Qwen3.5-2B-GGUF |
| UD-Q5_K_XL | 1.4 GBfile size | 2.0 GB1.4 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:UD-Q5_K_XLOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:UD-Q5_K_XLDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-UD-Q5_K_XL.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q6_K | 1.5 GBfile size | 2.1 GB1.5 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q6_KOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q6_KDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q6_K.gguf" --local-dir Qwen3.5-2B-GGUF |
| UD-Q6_K_XL | 1.7 GBfile size | 2.3 GB1.7 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:UD-Q6_K_XLOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:UD-Q6_K_XLDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-UD-Q6_K_XL.gguf" --local-dir Qwen3.5-2B-GGUF |
| Q8_0 | 1.9 GBfile size | 2.5 GB1.9 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:Q8_0Ollama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:Q8_0Download hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-Q8_0.gguf" --local-dir Qwen3.5-2B-GGUF |
| UD-Q8_K_XL | 2.6 GBfile size | 3.2 GB2.6 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:UD-Q8_K_XLOllama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:UD-Q8_K_XLDownload hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-UD-Q8_K_XL.gguf" --local-dir Qwen3.5-2B-GGUF |
| BF16 | 3.5 GBfile size | 4.1 GB3.5 + 0.09 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.5-2B-GGUF:BF16Ollama ollama run hf.co/unsloth/Qwen3.5-2B-GGUF:BF16Download hf download unsloth/Qwen3.5-2B-GGUF --include "Qwen3.5-2B-BF16.gguf" --local-dir Qwen3.5-2B-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLX 4bitlmstudio-community/Qwen3.5-2B-MLX-4bit | 1.6 GBfile size | 2.2 GB1.6 + 0.09 + 0.50 | Download page |
| MLX 4bitmlx-community/Qwen3.5-2B-MLX-4bit | 1.6 GBfile size | 2.2 GB1.6 + 0.09 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| BF16original weights | 4.2 GBfile size | 4.8 GB4.2 + 0.09 + 0.50 |
hf download Qwen/Qwen3.5-2B --local-dir Qwen3.5-2BOther conversions (11; sizes not read)
- litert-community/Qwen3.5-2B Other, 5.5k downloads
- onnx-community/Qwen3.5-2B-ONNX-OPT Other, 1.9k downloads
- taobao-mnn/Qwen3.5-2B-MNN Other, 1.4k downloads
- HanzoHuang/Qwen3.5-2B-RKLLM Other, 1.3k downloads
- cyankiwi/Qwen3.5-2B-AWQ-4bit AWQ, 4bit, 1.2k downloads
- AxionML/Qwen3.5-2B-NVFP4 NVFP4, 396 downloads
- axteam/Qwen3.5-2B-AWQ-4bit AWQ, 4bit, 259 downloads
- axteam/Qwen3.5-2B-FP8-dynamic FP8, 247 downloads
- axteam/Qwen3.5-2B-NVFP4A16 NVFP4, 229 downloads
- axteam/Qwen3.5-2B-AWQ-8bit AWQ, 8bit, 227 downloads
- axteam/Qwen3.5-2B-NVFP4 NVFP4, 226 downloads
The Hub tags 100 or more quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.
Run it by API
No provider on Hugging Face Inference Providers serves this model (as of 19 Nov 2026, 13:58 UTC). Other hosts may; we only list what the router reports.
Not built yet: one key for every open model, with BlockRun, and our own tested-quality score per provider. Until then, precision and measured speed are the best guide to a provider here.
Fine-tunes of this model
Repos that declare this model as their base, most downloaded first. Listing is not a recommendation; unreviewed ones have not been checked.
- unsloth/Qwen3.5-2B2.3B127kUnreviewed
- Alibaba-DAMO-Academy/RynnBrain1.1-2B2.2B106kUnreviewed
- openbmb/MiniCPM-OCR-Jev2.2B72kUnreviewed
- AlioLeuchtmann/qwen-3.5-sql-27B-Distill-2B15kUnreviewed
- prism-ml/Ternary-Bonsai-2-2B-mlx-2bit10kUnreviewed
- huihui-ai/Huihui-Qwen3.5-2B-abliterated2.3B3.5kUnreviewed
- internlm/Intern-Decision-2B2.2B2.0kUnreviewed
- THUSI-Lab/GameScaling-2B-S32.7B1.2kUnreviewed
- moganai/MoganOCR-TR2.2B663Unreviewed
- THUSI-Lab/GameScaling-2B-S12.7B440Unreviewed
- Cactus-Compute/Qwen3.5-2B22Unreviewed
Planned for this page Not built yet
- Try it in private chat
- A link to Solana AI's confidential chat for the models it serves. This model is not listed as served there.
- Fine-tune it
- Fine-tuning a few base models from this page and publishing the result here.
- Creators and earnings
- Pages for people who publish fine-tunes and adapters of this model, and what they earn.
- Fund and bounties
- Bounties for new open models, paid from a public fund. How the fund will work.
- Contribute compute
- How to give compute to open model work, with Akash.
Model data from the Hugging Face Hub 19 Nov 2026, 13:43 UTC. The repo. Found a mistake on this page? Tell us.