
DeepSeek-V4.1-Flash-20260910
DeepSeek-V4.1-Flash-20260910 is an open-weight model from DeepSeek, released Oct 17, 2026. It ranks #37 of 351 on the LMArena text leaderboard with a rating of 1464, and #9 of 27 on the Vals Index with 61.26%. It costs $0.15 per million input tokens and $0.60 per million output tokens, with a 1M-token context window.
- Developer
- DeepSeek
- Released
- Oct 17, 2026
- Context window
- 1M
- Price in / out, per 1M tokens
- $0.15 / $0.60
- Weights
- Open · MIT
- Inputs
- Text, Images
- Reasoning
- Yes
- Tool use
- Yes
- Hugging Face downloads, 30 days
- 73.2K
- Hugging Face likes
- 2.8K
Benchmarks
| Benchmark | Score bar | Accuracy | Rank | Cost / test |
|---|---|---|---|---|
| Vals IndexVals AI | 61.26% | 9 / 27 | $0.66 | |
| Terminal-Bench 4.0Coding | 19.70% | 30 / 55 | $0.50 | |
| Vibe Code BenchCoding | 84.74% | 18 / 115 | $0.41 | |
| Code MigrationCoding | 58.94% | 9 / 28 | $1.09 | |
| Finance AgentFinance | 62.93% | 12 / 31 | $0.47 | |
| Finance IndexFinance | 57.08% | 16 / 31 | $0.25 | |
| Tax Agent BenchFinance | 26.63% | 36 / 76 | $0.13 | |
| Legal IndexLaw | 71.86% | 9 / 28 | $0.59 | |
| Legal Research BenchLaw | 41.35% | 27 / 85 | $0.25 | |
| MedCodeMedicine | 41.17% | 57 / 107 | $0.3 / $1.2 | |
| Terminal-Bench ScienceReasoning | 0.00% | 41 / 42 | $0.47 |
Vals AI, as of Oct 24, 2026.
LMArena, by category
| Leaderboard | Rating | Rank | Votes |
|---|---|---|---|
| TextLMArena | 1464 | 37 / 351 | 13.8K |
| CodingLMArena | 1501 | 22 / 346 | 3.8K |
| MathLMArena | 1479 | 37 / 339 | 733 |
| Creative writingLMArena | 1439 | 47 / 349 | 2.8K |
| Instruction followingLMArena | 1474 | 28 / 351 | 4.8K |
| Hard promptsLMArena | 1483 | 33 / 351 | 8.8K |
| Multi-turnLMArena | 1460 | 47 / 349 | 2K |
| Long queriesLMArena | 1474 | 39 / 330 | 6K |
| ExpertLMArena | 1502 | 27 / 305 | 1.5K |
| Legal and governmentIndustries | 1445 | 79 / 325 | 1.2K |
| Medicine and healthcareIndustries | 1483 | 23 / 325 | 938 |
| Business and financeIndustries | 1461 | 31 / 344 | 2.4K |
| ChineseLanguages | 1501 | 51 / 328 | 1K |
| SpanishLanguages | 1458 | 42 / 249 | 477 |
| GermanLanguages | 1468 | 33 / 261 | 246 |
| FrenchLanguages | 1467 | 53 / 252 | 539 |
| JapaneseLanguages | 1428 | 45 / 236 | 295 |
| WebDevLMArena | 1665 | 14 / 124 | 2.5K |
| VisionLMArena | 1278 | 38 / 146 | 4.1K |
LMArena, CC BY 4.0, as of Oct 22, 2026. Where a model is rated at several effort levels, its best is shown (DeepSeek-V4.1-Flash-20260910 (Max), DeepSeek-V4.1-Flash-20261017 (Max)).
Compare DeepSeek-V4.1-Flash-20260910
DeepSeek
A research lab known for strong open-weight reasoning and coding models trained at low cost. Founded 2023, based in Hangzhou. deepseek.com
Questions
Who makes DeepSeek-V4.1-Flash-20260910?
DeepSeek-V4.1-Flash-20260910 is made by DeepSeek, founded 2023 and based in Hangzhou.
How much does DeepSeek-V4.1-Flash-20260910 cost?
$0.15 per million input tokens and $0.60 per million output tokens, at its provider’s list price.
What is DeepSeek-V4.1-Flash-20260910’s context window?
1M tokens (1,048,576).
How does DeepSeek-V4.1-Flash-20260910 rank?
#37 of 351 on LMArena’s text leaderboard, rated 1464 (95% interval 1458–1470) from 13,750 votes, as of Oct 22, 2026.
Is DeepSeek-V4.1-Flash-20260910 open source?
DeepSeek-V4.1-Flash-20260910 has open weights under the MIT license, downloadable from Hugging Face (deepseek-ai/DeepSeek-V4.1-Flash-1017).
Gemini 4 Argon
Claude Opus 5.5
Muse Wonder
MiMo-V2.6-Pro
Qwen3.8-Max
Kimi K3
GLM-5.3
GPT-5.5
ERNIE 5.1
Grok 4.20 Beta
amazon-nova-experimental-chat
Dola Seed 2.0 Pro
NVIDIA-Nemotron-3-Ultra-550B-A55B
Gemma 4 31B
Step 5 Preview