- Size
- 763B parameterscounted from the safetensors header on the Hub
- Licence
- unknown, MIT LicenseOur reading, check the licence: You can use, change and redistribute the model for any purpose, including commercially. Keep the copyright notice and the licence text with copies. You do not have to share your changes.
- On the Hub since
- 17 Oct 2026the date the repo was created
- Context length
- 1M tokensfrom the model's config
- Kind
- Base model (no base model declared on the Hub)DeepseekV41ForCausalLM
- On Hugging Face
- deepseek-ai/DeepSeek-V4.1-Flash-1017502k downloads in the last 30 days, 6.0k likes
Scores
- effort max46.0%±5.019 Oct 2026, log
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: claude-fable-5-2_max, 79.0% (Anthropic). All 166 results
- effort max46.3%±7.919 Oct 2026, Epoch AI
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: claude-fable-5-2_max, 100.0% (Anthropic). All 83 results
- effort max71.2%±2.719 Oct 2026, Epoch AI
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: claude-fable-5-2_max, 95.1% (Anthropic). All 127 results
- effort max93.2%±1.619 Oct 2026, log
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: gpt-6.2-astra_max, 96.8% (OpenAI). All 341 results
- effort max49.0%±5.019 Oct 2026, Epoch AI
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: gpt-6.2-astra_max, 95.0% (OpenAI). All 148 results
- effort max99.4%±0.419 Oct 2026, log
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: claude-fable-5_high, 100.0% (Anthropic). All 318 results
- effort max42.5%±1.619 Oct 2026, log
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: claude-fable-5-2_max, 94.0% (Anthropic). All 105 results
- effort max69.8%via Epoch AI (error bar left out)
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: gemini-4-carbon_unknown, 84.2% (Google DeepMind). All 66 results
- effort high94.5%arcprize.org
- effort max93.5%arcprize.org
- effort low91.5%arcprize.org
- effort none5.7%arcprize.org
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: claude-fable-5_max, 98.5% (Anthropic). All 318 results
- effort max78.1%via Epoch AI
- effort low73.6%via Epoch AI
- effort high68.6%via Epoch AI
- effort none2.1%via Epoch AI
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: gpt-6-astra_max, 95.0% (OpenAI). All 299 results
- effort max24.0%via Epoch AI
- effort none0.3%via Epoch AI
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: gemini-4-carbon_high, 36.0% (Google DeepMind). All 271 results
- effort max; Reasoning level: Max47.7%cursor.com
- effort medium; Reasoning level: Medium46.3%cursor.com
- effort high; Reasoning level: High45.6%cursor.com
- effort low; Reasoning level: Low44.9%cursor.com
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: gemini-4-carbon_high, 60.4% (Google DeepMind). All 94 results
- effort high; Agent: codex_cli; Reasoning effort: high19.7%htihle.github.io
Epoch's name: deepseek-v4.1-flash-1017. Best closed model on this set: claude-fable-5-2_xhigh, 65.3% (Anthropic). All 27 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: 0.63 GB, worked out from the model's config (deepseek-ai/DeepSeek-V4.1-Flash-1017), 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/DeepSeek-V4.1-Flash-1017-GGUF (converted by unsloth), 1.6M 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 (0.90 GB), not counted below.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| UD-IQ1_S | 217.0 GBfile size, 5 parts | 218.1 GB217.0 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-IQ1_SDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-IQ1_S/DeepSeek-V4.1-Flash-1017-UD-IQ1_S-*-of-00005.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-IQ1_M | 233.2 GBfile size, 5 parts | 234.3 GB233.2 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-IQ1_MDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-IQ1_M/DeepSeek-V4.1-Flash-1017-UD-IQ1_M-*-of-00005.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-IQ2_XXS | 279.8 GBfile size, 6 parts | 281.0 GB279.8 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-IQ2_XXSDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-IQ2_XXS/DeepSeek-V4.1-Flash-1017-UD-IQ2_XXS-*-of-00006.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-Q2_K_XL | 311.8 GBfile size, 7 parts | 312.9 GB311.8 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-Q2_K_XLDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-Q2_K_XL/DeepSeek-V4.1-Flash-1017-UD-Q2_K_XL-*-of-00007.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-IQ3_XXS | 367.6 GBfile size, 7 parts | 368.8 GB367.6 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-IQ3_XXSDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-IQ3_XXS/DeepSeek-V4.1-Flash-1017-UD-IQ3_XXS-*-of-00007.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-IQ3_S | 394.5 GBfile size, 8 parts | 395.7 GB394.5 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-IQ3_SDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-IQ3_S/DeepSeek-V4.1-Flash-1017-UD-IQ3_S-*-of-00008.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-Q3_K_XL | 451.8 GBfile size, 8 parts | 452.9 GB451.8 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-Q3_K_XLDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-Q3_K_XL/DeepSeek-V4.1-Flash-1017-UD-Q3_K_XL-*-of-00008.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-Q4_K_XL | 466.3 GBfile size, 9 parts | 467.4 GB466.3 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-Q4_K_XLDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-Q4_K_XL/DeepSeek-V4.1-Flash-1017-UD-Q4_K_XL-*-of-00009.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
| UD-Q8_K_XL | 473.5 GBfile size, 9 parts | 474.6 GB473.5 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF:UD-Q8_K_XLDownload hf download unsloth/DeepSeek-V4.1-Flash-1017-GGUF --include "UD-Q8_K_XL/DeepSeek-V4.1-Flash-1017-UD-Q8_K_XL-*-of-00009.gguf" --local-dir DeepSeek-V4.1-Flash-1017-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLXddalcu/DeepSeek-V4.1-Flash-1017-MLX-Serve-iQ-MLX-3.0bpw | 397.9 GBfile size | 399.1 GB397.9 + 0.63 + 0.50 | Download page |
| MLXmlx-community/DeepSeek-V4.1-Flash-1017-oQ4e-mtp | 402.4 GBfile size | 403.5 GB402.4 + 0.63 + 0.50 | Download page |
Other conversions (14; sizes not read)
- ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF GGUF, 4.4M downloads; sizes not used: file sizes do not match this model's 763.2B parameters (median 0.26 times the nominal size); the repo may be tagged wrongly
- ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-Coder-GGUF GGUF, 523k downloads; sizes not used: file sizes do not match this model's 763.2B parameters (median 0.26 times the nominal size); the repo may be tagged wrongly
- nvidia/DeepSeek-V4.1-Flash-1017-NVFP4 NVFP4, 72k downloads
- bertholomus/DeepSeek-V4.1-Flash-1017-EXL3-2.5bpw EXL3, 26k downloads
- Mia-AiLab/DeepSeek-V4.1-Flash-1017-EXL3-EAQ-v1-2.9-bpw EXL3, 6.8k downloads
- hyukudan/DeepSeek-V4.1-Flash-1017-strix-trellis-EXL3 EXL3, 6.4k downloads
- LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw EXL3, 5.1k downloads
- beamster/DeepSeek-V4.1-Flash-1017-Sushi-STQ-1.75bpw FP8, 4.5k downloads
- immiq/DeepSeek-V4.1-Flash-1017-Orinfer-A2-W4A8 Other, 2.9k downloads
- tiny-random/deepseek-v4.1-flash-1017 FP8, 1.4k downloads
- Blackfrost-AI/DeepSeek-V4.1-Flash-1017-DERISKED-FP8 FP8, 1.1k downloads
- LibertAIDAI/DeepSeek-V4.1-Flash-1017-NVFP4 NVFP4, 914 downloads
- ewhacc/DeepSeek-V4.1-Flash-1017-Tandem-VQ2 FP8, 735 downloads
- brandonmusic/DeepSeek-V4.1-Flash-1017-TrellisMX-K2-Mix Other, 589 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
35 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 |
|---|---|---|---|---|---|---|---|---|---|
| ovhcloud | $0 | $0 | 1M | not given | 98 | 0.5 s | ? | yes | yes |
| DeepInfra | $0.098 | $0.29 | 1M | fp8 | 106 | 0.9 s | 98.5% | yes | yes |
| Pearl Research | $0.05 | $0.30 | 1M | fp8 | 136 | 1.1 s | 90.3% | yes | yes |
| StepFun | $0.10 | $0.42 | 977k | fp8 | 130 | 1.0 s | 100.0% | yes | yes |
| Ant Digital | $0.12 | $0.48 | 1M | fp8 | 85 | 1.6 s | 100.0% | yes | no |
| Novita | $0.12 | $0.48 | 1M | fp8 | 116 | 1.3 s | 100.0% | yes | yes |
| River | $0.12 | $0.48 | 1M | not given | 95 | 1.0 s | 100.0% | yes | yes |
| Cohere | $0.15 | $0.50 | 1M | fp8 | 149 | 0.8 s | 100.0% | yes | yes |
| Particle AI | $0.06 | $0.50 | 1M | not given | 121 | 0.5 s | 100.0% | yes | yes |
| Relace | $0.12 | $0.50 | 1M | fp4 | 143 | 0.9 s | 100.0% | yes | yes |
| Wafer | $0.036 | $0.50 | 1M | mxfp4 | 133 | 0.4 s | 100.0% | yes | yes |
| Morph | $0.059 | $0.52 | 1M | fp8 | 115 | 0.6 s | 99.6% | yes | yes |
| Cloudflare | $0.10 | $0.60 | 1M | not given | 101 | 1.9 s | 99.7% | yes | yes |
| DekaLLM | $0.10 | $0.60 | 1M | not given | 143 | 1.7 s | 100.0% | yes | yes |
| Inceptron | $0.10 | $0.60 | 1M | fp4 | 139 | 0.4 s | 99.6% | yes | yes |
| SiliconFlow | $0.15 | $0.60 | 1M | fp8 | 135 | 1.1 s | 99.9% | yes | yes |
| Ionstream | $0.21 | $0.62 | 1M | not given | 192 | 0.8 s | 99.1% | yes | yes |
| InferenceNet | $0.08 | $0.64 | 1016k | fp8 | 268 | 0.2 s | 100.0% | yes | yes |
| ModelRun | $0.17 | $0.67 | 1M | fp4 | 112 | 0.9 s | 100.0% | yes | yes |
| PrimeIntellect | $0.044 | $0.70 | 1M | not given | 112 | 0.6 s | 100.0% | yes | yes |
| DigitalOcean | $0.18 | $0.72 | 1M | not given | 146 | 0.8 s | 100.0% | yes | yes |
| Nebius | $0.18 | $0.72 | 1M | fp4 | 240 | 0.5 s | 100.0% | yes | yes |
| Makora | $0.15 | $0.80 | 1M | fp8 | 154 | 1.1 s | 97.3% | yes | yes |
| Fireworks | $0.21 | $0.84 | 1M | not given | 115 | 1.0 s | 76.6% | yes | yes |
| OpenInference | $0.06 | $0.90 | 1M | fp4 | 146 | 0.9 s | 100.0% | yes | yes |
| Mistral | $0.24 | $0.96 | 1M | fp8 | 154 | 1.1 s | 100.0% | yes | yes |
| PrimaLabs | $0.24 | $0.96 | 1M | not given | 126 | 1.3 s | 99.7% | yes | yes |
| Modal | $0.25 | $1.00 | 1M | not given | 104 | 0.9 s | 100.0% | yes | yes |
| BaseTen | $0.30 | $1.20 | 1M | not given | 133 | 0.7 s | 100.0% | yes | no |
| CoreWeave | $0.30 | $1.20 | 1M | fp8 | 217 | 0.5 s | 100.0% | yes | yes |
| Crusoe | $0.29 | $1.20 | 1M | fp8 | 210 | 0.9 s | 98.2% | yes | yes |
| Near AI | $0.30 | $1.20 | 1M | not given | 43 | 0.7 s | 99.3% | yes | yes |
| Parasail | $0.30 | $1.20 | 1M | fp8 | 202 | 0.5 s | 100.0% | yes | yes |
| Together | $0.30 | $1.20 | 1M | not given | 213 | 0.4 s | 100.0% | yes | yes |
| Venice | $0.30 | $1.20 | 977k | fp8 | 135 | 1.4 s | 100.0% | yes | yes |
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": "deepseek-ai/DeepSeek-V4.1-Flash-1017:ovhcloud", "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.
- OBLITERATUS/DeepSeek-V4.1-Flash-1017-obliterated755B21kUnreviewed
- dealignai/DeepSeek-V4.1-Flash-1017-UNCENSORED-FP8763B16kUnreviewed
- apetersson/DeepSeek-V4.1-Flash-1017-Abliterated-MLX14kUnreviewed
- pentacoxian-dev/DeepSeek-V4.1-1017-Ablitrated-Efficient-v3763B306Unreviewed
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.