- Size
- 36.0B parameters, 3.0B active (from the repo name)counted 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
- 15 Apr 2026the date the repo was created
- Context length
- 256k tokensfrom the model's config
- Kind
- Base model (no base model declared on the Hub)Qwen3_5MoeForConditionalGeneration
- On Hugging Face
- Qwen/Qwen3.6-35B-A3B3.9M downloads in the last 30 days, 3.2k likes
Scores
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: claude-fable-5-2_max, 95.1% (Anthropic). All 127 results
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: gpt-6.2-astra_max, 96.8% (OpenAI). All 341 results
- effort none22.0%±4.227 Aug 2026, log
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: gpt-6.2-astra_max, 95.0% (OpenAI). All 148 results
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: claude-fable-5_high, 100.0% (Anthropic). All 318 results
- default setting0.3%via Epoch AI
- effort none0.0%via Epoch AI
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: gemini-4-carbon_high, 36.0% (Google DeepMind). All 271 results
- default setting73.9%conceptualreasoning.ai
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: claude-fable-5-2_unknown, 99.2% (Anthropic). All 241 results
- default setting29.7%25.0 to 34.4conceptualreasoning.ai
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: claude-fable-5-2_xhigh, 70.6% (Anthropic). All 203 results
- default setting44.4%yhenon.github.io
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: claude-fable-5_high, 95.0% (Anthropic). All 29 results
- Agent: little-coder24.6%±3.216 Apr 2026, tbench.ai
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: gpt-5.5_unknown, 84.7% (OpenAI). All 202 results
- default setting34.5%±0.0htihle.github.io
Epoch's name: qwen3.6-35b-a3b. Best closed model on this set: gpt-6-astra_promax, 93.6% (OpenAI). All 173 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: 0.16 GB, worked out from the model's config (Qwen/Qwen3.6-35B-A3B), 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/Qwen3.6-35B-A3B-MTP-GGUF (converted by unsloth), 2.7M 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 (0.84 GB), not counted below.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| UD-IQ3_XXS | 13.1 GBfile size | 13.8 GB13.1 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ3_XXSOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ3_XXSDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-IQ3_XXS.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-IQ3_S | 14.3 GBfile size | 14.9 GB14.3 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ3_SOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ3_SDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-IQ3_S.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q3_K_M | 15.9 GBfile size | 16.6 GB15.9 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q3_K_MOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q3_K_MDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q3_K_M.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q3_K_XL | 16.0 GBfile size | 16.7 GB16.0 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q3_K_XLOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q3_K_XLDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q3_K_XL.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-IQ4_XS | 17.0 GBfile size | 17.6 GB17.0 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ4_XSOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ4_XSDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-IQ4_XS.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-IQ4_NL | 17.3 GBfile size | 17.9 GB17.3 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ4_NLOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-IQ4_NLDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-IQ4_NL.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q4_K_S | 19.9 GBfile size | 20.6 GB19.9 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q4_K_SOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q4_K_SDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q4_K_S.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| MXFP4_MOE | 20.7 GBfile size | 21.3 GB20.7 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:MXFP4_MOEOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:MXFP4_MOEDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-MXFP4_MOE.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q4_K_M | 21.1 GBfile size | 21.8 GB21.1 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q4_K_MOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q4_K_MDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q4_K_M.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q4_K_XL | 21.3 GBfile size | 21.9 GB21.3 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q4_K_XLOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q4_K_XLDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q4_K_XL.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q5_K_S | 23.8 GBfile size | 24.4 GB23.8 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q5_K_SOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q5_K_SDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q5_K_S.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q5_K_M | 25.2 GBfile size | 25.9 GB25.2 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q5_K_MOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q5_K_MDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q5_K_M.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q5_K_XL | 25.3 GBfile size | 26.0 GB25.3 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q5_K_XLOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q5_K_XLDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q5_K_XL.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q6_K | 28.0 GBfile size | 28.6 GB28.0 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q6_KOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q6_KDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q6_K.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q6_K_XL | 30.4 GBfile size | 31.0 GB30.4 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q6_K_XLOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q6_K_XLDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q6_K_XL.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| Q8_0 | 35.2 GBfile size | 35.9 GB35.2 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:Q8_0Ollama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:Q8_0Download hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-Q8_0.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| UD-Q8_K_XL | 36.4 GBfile size | 37.1 GB36.4 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q8_K_XLOllama ollama run hf.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF:UD-Q8_K_XLDownload hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "Qwen3.6-35B-A3B-UD-Q8_K_XL.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
| BF16 | 66.2 GBfile size, 2 parts | 66.8 GB66.2 + 0.16 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Qwen3.6-35B-A3B-MTP-GGUF:BF16Download hf download unsloth/Qwen3.6-35B-A3B-MTP-GGUF --include "BF16/Qwen3.6-35B-A3B-BF16-*-of-00002.gguf" --local-dir Qwen3.6-35B-A3B-MTP-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLX 4bitlmstudio-community/Qwen3.6-35B-A3B-MLX-4bit | 19.0 GBfile size | 19.7 GB19.0 + 0.16 + 0.50 | Download page |
| MLX 8bitlmstudio-community/Qwen3.6-35B-A3B-MLX-8bit | 35.1 GBfile size | 35.8 GB35.1 + 0.16 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| BF16original weights | 67.0 GBfile size | 67.6 GB67.0 + 0.16 + 0.50 |
hf download Qwen/Qwen3.6-35B-A3B --local-dir Qwen3.6-35B-A3BOther conversions (16; sizes not read)
- Qwen/Qwen3.6-35B-A3B-FP8 FP8, by the maker, 3.8M downloads
- nvidia/Qwen3.6-35B-A3B-NVFP4 NVFP4, 4.1M downloads
- unsloth/Qwen3.6-35B-A3B-NVFP4 NVFP4, 774k downloads
- RedHatAI/Qwen3.6-35B-A3B-NVFP4 NVFP4, 739k downloads
- unsloth/Qwen3.6-35B-A3B-NVFP4-Fast NVFP4, 687k downloads
- QuantTrio/Qwen3.6-35B-A3B-AWQ AWQ, 148k downloads
- cyankiwi/Qwen3.6-35B-A3B-AWQ-4bit AWQ, 4bit, 115k downloads
- palmfuture/Qwen3.6-35B-A3B-GPTQ-Int4 GPTQ, int4, 88k downloads
- lokeshe09/Qwen3.6-35B-A3B-INT8 Other, int8, 77k downloads
- neroued/Qwen3.6-35B-A3B-NInfer Other, 45k downloads
- RedHatAI/Qwen3.6-35B-A3B-FP8 FP8, 13k downloads
- btbtyler09/Qwen3.6-35B-A3B-GPTQ-8bit GPTQ, 8bit, 11k downloads
- btbtyler09/Qwen3.6-35B-A3B-GPTQ-4bit GPTQ, 4bit, 3.2k downloads
- Intel/Qwen3.6-35B-A3B-int4-mixed-AutoRound Other, int4, 2.0k downloads
- RedHatAI/Qwen3.6-35B-A3B-FP8-dynamic FP8, 1.3k downloads
- amd/Qwen3.6-35B-A3B-w4a16-llmcompressor Other, w4a16, 1.2k 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
7 providers serve this model through the Hugging Face router. Prices are US dollars per million tokens. Precision is what the provider reports; a lower precision than the original weights can change answers. Live first, then cheapest output.
| Provider | In | Out | Context | Precision | Tokens/s | First token | Uptime 24h | Tools | JSON |
|---|---|---|---|---|---|---|---|---|---|
| CoreWeave | $0.10 | $0.89 | 256k | fp8 | 118 | 0.5 s | 100.0% | no | yes |
| AkashML | $0.10 | $0.90 | 256k | fp8 | 38 | 1.4 s | 100.0% | yes | yes |
| Parasail | $0.10 | $0.90 | 256k | fp8 | 42 | 0.8 s | 100.0% | yes | yes |
| DeepInfra | $0.10 | $0.95 | 256k | fp8 | 27 | 0.6 s | 99.5% | yes | yes |
| Venice | $0.10 | $1.00 | 250k | fp8 | 27 | 0.7 s | 99.8% | yes | no |
| PrimeIntellect | $0.11 | $1.01 | 256k | fp8 | 92 | 0.4 s | 100.0% | yes | yes |
| featherless-ai | ? | ? | ? | not given | ? | ? | ? | not given | not given |
Through the router, with a Hugging Face token (the router bills your Hugging Face account). Change the name after the colon to pick another provider.
curl https://router.huggingface.co/v1/chat/completions \
-H "Authorization: Bearer $HF_TOKEN" -H "Content-Type: application/json" \
-d '{"model": "Qwen/Qwen3.6-35B-A3B:coreweave", "messages": [{"role": "user", "content": "Hello"}]}'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.
- Bahushruth/Qwen3.6-35B-A3B-abliterated-v434.7B278kUnreviewed
- MemTensor/MemT-1.0-Nano35.1B90kUnreviewed
- incoai/Qwen3.6-35B-A3B-DFlash289kUnreviewed
- Diffuspace/Clover-35B-A3B-Preview34.7B5.9kUnreviewed
- unsloth/Qwen3.6-35B-A3B36.0B5.8kUnreviewed
- endless-frontier/Fx-Work-102836.0B4.3kUnreviewed
- empero-ai/Qwen3.8-35B-A3B-Distill35.1B4.3kUnreviewed
- endless-frontier/Fx-Phys36.0B4.2kUnreviewed
- huihui-ai/Huihui-Qwen3.6-35B-A3B-abliterated36.0B3.2kUnreviewed
- mlx-community/Qwen3.6-35B-A3B-bf1635.1B2.9kUnreviewed
- Apexintelligence-AI/Apex-35B-V1.036.0B2.4kUnreviewed
- Accio-Lab/occamy-1.035.1B1.6kUnreviewed
- whittleOrg/Whittle-Next-27B-A3B27.6B1.4kUnreviewed
- oyildirim/CyberStrike-OffSec-35B35.1B700Unreviewed
- mercor/Qwen3.6-35B-A3B-Mercor35.1B54Unreviewed
- PhaiModelLab/Phai-A1-35B-1M35.1B33Unreviewed
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.