DeepSeek
DeepSeek V4.1 Flash benchmark results
6 results on 5 benchmarks from 2 sources.
Also listed as: DeepSeek-V4.1-Flash-20260910 (Max) · DeepSeek-V4.1-Flash-20261017 (Max) · deepseek-v4.1-flash_max · DeepSeek V4.1 Flash (max)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1476 | 34 | Overall, style control on; 13750 votes; 95% CI 1470–1482“DeepSeek-V4.1-Flash-20260910 (Max)” | Official leaderboard | Oct 22, 2026 |
| LMArena WebDev | 1664 | 13 | Overall; 2425 votes; 95% CI 1651–1677“DeepSeek-V4.1-Flash-20261017 (Max)” | Official leaderboard | Oct 22, 2026 |
| LMArena WebDev | 1618 | 26 | Overall; 3526 votes; 95% CI 1607–1629“DeepSeek-V4.1-Flash-20260910 (Max)” | Official leaderboard | Oct 22, 2026 |
| LMArena Vision | 1261 | 46 | Overall, style control on; 4100 votes; 95% CI 1251–1272“DeepSeek-V4.1-Flash-20260910 (Max)” | Official leaderboard | Oct 19, 2026 |
| GPQA Diamond | 89.8% | 45 | Epoch AI run, accuracy; effort: max | Independent run | Oct 9, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 98.3% | 25 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: max | Independent run | Oct 8, 2026 |
Where DeepSeek V4.1 Flash ranks
Primary setting, one best-sourced score per model (official / independent first); neighbours in brackets.
| Benchmark | Score | Rank | Nearby models | #1 |
|---|---|---|---|---|
| LMArena Text | 1476 | #29 / 392 | Gemini 3.5 Flash 1477 · GPT-6 Astra 1477 · GLM-5.3-Flash 1476 · GPT-5.2 Chat (20260210) 1476 | Gemini 4 Argon1517 |
| LMArena WebDev | 1664 | #12 / 135 | Gemini 4 Argon 1673 · Qwen 3.8 Max 1673 · Muse Spark 1.3 1657 · Kimi K3 1655 | Claude Opus 5.51807 |
| AIME (OTIS Mock AIME 2024–2025) | 98.3% | #25 / 200 | DeepSeek V4 Pro 0813 98.6% · Mistral Large 4 98.3% · GPT-5.6 Luna 98.3% · Muse Spark Open 98.1% | Gemini 4 Argon100.0% |
| GPQA Diamond | 89.8% | #39 / 217 | Claude Opus 4.7 90.2% · GLM-5.1 89.9% · Muse Spark 89.8% · Claude Haiku 5.5 89.7% | Gemini 4 Argon96.4% |
| LMArena Vision | 1261 | #39 / 154 | Claude Sonnet 5 1262 · Gemma 4 31B 1262 · GPT-5.6 Luna 1257 · Dola Seed 2.0 Pro 1257 | Claude Fable 51309 |
Other DeepSeek models
DeepSeek-V3.2DeepSeek-V3DeepSeek-R1 (May 2025)DeepSeek-V3 (Mar 2025)DeepSeek LLM 67BDeepSeek-R1DeepSeek-V4-ProDeepSeek-R1-Distill-Llama-70BDeepSeek-R1-Distill-Qwen-14B
Related on this site
FAQ
Where does DeepSeek V4.1 Flash rank best?
LMArena Text: #29 of 392 (1476); #1 there is Gemini 4 Argon (1517).
Where do these DeepSeek V4.1 Flash scores come from?
From Epoch AI Benchmarking Hub and LMArena leaderboard dataset. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.