- Size
- 4.0B 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
- 27 Apr 2025the date the repo was created
- Context length
- 40k tokensfrom the model's config
- Fine-tune
- of Qwen/Qwen3-4B-BaseQwen3ForCausalLM
- On Hugging Face
- Qwen/Qwen3-4B3.1M downloads in the last 30 days, 804 likes
Scores
Epoch's name: qwen3-4b. Best closed model on this set: gpt-6.2-astra_max, 96.8% (OpenAI). All 341 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: 1.1 GB, worked out from the model's config (Qwen/Qwen3-4B), 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
Qwen/Qwen3-4B-GGUF (the maker's own conversion), 757k downloads. Sizes are the real files on the Hub, all parts added up.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| Q4_K_M | 2.3 GBfile size | 4.0 GB2.3 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-4B-GGUF:Q4_K_MOllama ollama run hf.co/Qwen/Qwen3-4B-GGUF:Q4_K_MDownload hf download Qwen/Qwen3-4B-GGUF --include "Qwen3-4B-Q4_K_M.gguf" --local-dir Qwen3-4B-GGUF |
| Q5_0 | 2.6 GBfile size | 4.3 GB2.6 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-4B-GGUF:Q5_0Ollama ollama run hf.co/Qwen/Qwen3-4B-GGUF:Q5_0Download hf download Qwen/Qwen3-4B-GGUF --include "Qwen3-4B-Q5_0.gguf" --local-dir Qwen3-4B-GGUF |
| Q5_K_M | 2.7 GBfile size | 4.3 GB2.7 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-4B-GGUF:Q5_K_MOllama ollama run hf.co/Qwen/Qwen3-4B-GGUF:Q5_K_MDownload hf download Qwen/Qwen3-4B-GGUF --include "Qwen3-4B-Q5_K_M.gguf" --local-dir Qwen3-4B-GGUF |
| Q6_K | 3.1 GBfile size | 4.7 GB3.1 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-4B-GGUF:Q6_KOllama ollama run hf.co/Qwen/Qwen3-4B-GGUF:Q6_KDownload hf download Qwen/Qwen3-4B-GGUF --include "Qwen3-4B-Q6_K.gguf" --local-dir Qwen3-4B-GGUF |
| Q8_0 | 4.0 GBfile size | 5.6 GB4.0 + 1.1 + 0.50 | Commandsllama.cpp llama-server -hf Qwen/Qwen3-4B-GGUF:Q8_0Ollama ollama run hf.co/Qwen/Qwen3-4B-GGUF:Q8_0Download hf download Qwen/Qwen3-4B-GGUF --include "Qwen3-4B-Q8_0.gguf" --local-dir Qwen3-4B-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLX 4bitlmstudio-community/Qwen3-4B-MLX-4bit | 2.1 GBfile size | 3.7 GB2.1 + 1.1 + 0.50 | Download page |
| MLX 4bitmlx-community/Qwen3-4B-4bit | 2.1 GBfile size | 3.7 GB2.1 + 1.1 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| BF16original weights | 7.5 GBfile size | 9.1 GB7.5 + 1.1 + 0.50 |
hf download Qwen/Qwen3-4B --local-dir Qwen3-4BOther conversions (11; sizes not read)
- Qwen/Qwen3-4B-AWQ AWQ, by the maker, 94k downloads
- Qwen/Qwen3-4B-FP8 FP8, by the maker, 52k downloads
- unsloth/Qwen3-4B-unsloth-bnb-4bit bitsandbytes, 4bit, 51k downloads
- mobilint/Qwen3-4B Other, 28k downloads
- taobao-mnn/Qwen3-4B-MNN Other, 25k downloads
- mlc-ai/Qwen3-4B-q4f16_1-MLC Other, 14k downloads
- litert-community/Qwen3-4B Other, 14k downloads
- unsloth/Qwen3-4B-bnb-4bit bitsandbytes, 4bit, 8.8k downloads
- qwen-community/Qwen3-4B-FP8 FP8, 1.2k downloads
- RedHatAI/Qwen3-4B-FP8-dynamic FP8, 903 downloads
- JunHowie/Qwen3-4B-GPTQ-Int4 GPTQ, int4, 214 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.
- Qwen/Qwen3Guard-Gen-4B4.4B19kUnreviewed
- YannQi/R-4B4.8B4.6kUnreviewed
- vllm-project/R-4B4.8B3.9kUnreviewed
- Qwen/Qwen3-4B-SafeRL4.0B3.8kUnreviewed
- gugujiji/novel-humanize-4b4.0B2.0kUnreviewed
- huihui-ai/Qwen3-4B-abliterated4.0B233Unreviewed
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.