LLMs TigerLive LLM benchmarks

deepseek-v4.1-flash-max: benchmark scores, price and ranking

deepseek-v4.1-flash-max from DeepSeek ranks #75 of 743 LLMs on the LLMs Tiger aggregate benchmark score (52.2), based on 15 benchmark results captured up to 2026-10-13.

Organisation
DeepSeek
Type
Fast / low-cost (flash) model
Open weights
No
Released
2026-09-09 (source: epoch)
Aggregate rank
#75 of 743 (score 52.2)

Relative to other models, deepseek-v4.1-flash-max is strongest on OTIS Mock AIME (#17 of 189) and weakest on LMArena Text Elo (#22 of 28).

Across the 15 benchmarks where at least ten models are scored, deepseek-v4.1-flash-max is in the top 10% on 1 and in the top half on 10.

deepseek-v4.1-flash-max benchmark results

deepseek-v4.1-flash-max benchmark scores, ranked against all models measured on each benchmark
BenchmarkScoreRankSourceCaptured
OTIS Mock AIME98.3%#17 of 189Epoch AI Benchmarking Hub2026-10-13
GPQA Diamond89.8%#36 of 275Epoch AI Benchmarking Hub2026-10-13
WebDev Arena1,614#13 of 67Epoch AI Benchmarking Hub2026-10-13
ARC-AGI-272.9%#17 of 87ARC Prize official leaderboard2026-10-13
Mystery Games43.0%#13 of 59Epoch AI Benchmarking Hub2026-10-13
CritPt14.3%#30 of 115Epoch AI Benchmarking Hub2026-10-13
SciCode51.9%#29 of 106Epoch AI Benchmarking Hub2026-10-13
ALE-Bench1,092#24 of 79Epoch AI Benchmarking Hub2026-10-13
FrontierMath T1-367.4%#24 of 68Epoch AI Benchmarking Hub2026-10-13
SimpleQA Verified42.9%#32 of 69Epoch AI Benchmarking Hub2026-10-13
LMArena Code Elo1,620#16 of 30LMArena2026-10-07
FrontierMath T426.8%#30 of 50Epoch AI Benchmarking Hub2026-10-13
GDP.pdf19.8%#15 of 25Epoch AI Benchmarking Hub2026-10-13
Furniture Assembly34.2%#19 of 28Epoch AI Benchmarking Hub2026-10-13
LMArena Text Elo1,474#22 of 28LMArena2026-10-05

About deepseek-v4.1-flash-max

deepseek-v4.1-flash-max is a smaller, lower-cost model from DeepSeek. It was released on 2026-09-09, according to Epoch AI. No API price is listed for it yet. We hold 15 benchmark results for it across 4 areas, and 1 is in the top tenth of its board.

Where deepseek-v4.1-flash-max stands

On our aggregate score, deepseek-v4.1-flash-max is #75 of 743 ranked models, in the top quarter. Its score is 52.2; the scale is centred on 50, so it sits 2.2 points above the typical model. That beats 90% of the ranked field. Among the 33 ranked DeepSeek models it is #4. The next one up is DeepSeek-V3.2-Exp (high) at 52.4. The next one down is DeepSeek-V2.5-1210 at 51.7.

deepseek-v4.1-flash-max, benchmark by benchmark

On OTIS Mock AIME, deepseek-v4.1-flash-max scores 98.3% and ranks #17 of 189. Only 16 models score higher, led by Claude Fable 5.1 (max) at 100.0%, which is 1.7 points ahead. It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so deepseek-v4.1-flash-max is 40.6 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

On GPQA Diamond, deepseek-v4.1-flash-max reaches 89.8%. It ranks #36 of 275, ahead of 87% of the field, and 5.9 points behind GPT-6 Astra (max) at 95.8%. The results either side are GLM-5.3-Flash (max) (90.2%) and Muse Spark (89.8%). It covers expert-level science questions. The board median is 71.7%, so deepseek-v4.1-flash-max is 18.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

deepseek-v4.1-flash-max manages 1,614 on WebDev Arena, good for #13 of 67. Claude Opus 5.5 (max) leads at 1,820, 206 points further up, and 81% of models sit below. The results either side are GLM-5.3 (max) (1,614) and Gemini 3.7 Flash (high) (1,587). It covers building web apps, crowd-voted (Elo). The board median is 1,433, so deepseek-v4.1-flash-max is 180 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

ARC-AGI-2 puts deepseek-v4.1-flash-max at 72.9%, #17 of 87. That beats 80% of the models tested, but 22.1 points short of GPT-6 Astra (max) (95.0%). The results either side are GPT-5.4 (xhigh) (74.0%) and gemini-3.5-flash-high (72.1%). It covers abstract reasoning, ARC Prize public eval set. The board median is 22.8%, so deepseek-v4.1-flash-max is 50.1 points above it. Source: ARC Prize official leaderboard, captured 2026-10-13.

On Mystery Games, deepseek-v4.1-flash-max reaches 43.0%. It ranks #13 of 59, ahead of 78% of the field, and 41.0 points behind GPT-6 Astra (max) at 84.0%. The results either side are DeepSeek V4 Pro 0813 (max) (43.0%) and Qwen3.8 Max (xhigh) (38.0%). It covers deductive reasoning puzzles. The board median is 23.0%, so deepseek-v4.1-flash-max is 20.0 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

deepseek-v4.1-flash-max manages 14.3% on CritPt, good for #30 of 115. GPT-5.6 Sol (max) leads at 32.3%, 18.0 points further up, and 74% of models sit below. The results either side are Muse Spark 1.1 (xhigh) (15.1%) and Gemini 3.7 Flash (high) (14.3%). It covers unpublished research-level physics problems. The board median is 2.9%, so deepseek-v4.1-flash-max is 11.4 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

deepseek-v4.1-flash-max lands below the middle on FrontierMath T4: 26.8%, #30 of 50, with 58% of the field scoring higher. GPT-6.1 Sol (max) leads at 100.0%; the closest result above is DeepSeek V4 Pro 0813 (max) at 26.8%. It covers research-level maths, tier 4. The board median is 30.5%, so deepseek-v4.1-flash-max is 3.7 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

deepseek-v4.1-flash-max lands below the middle on GDP.pdf: 19.8%, #15 of 25, with 56% of the field scoring higher. GPT-6 Astra (max) leads at 34.2%; the closest result above is Claude Opus 4.7 (max) at 21.0%. It covers economically valuable document tasks. The board median is 22.8%, so deepseek-v4.1-flash-max is 3.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

deepseek-v4.1-flash-max lands below the middle on Furniture Assembly: 34.2%, #19 of 28, with 64% of the field scoring higher. Claude Opus 5.5 (max) leads at 83.3%; the closest result above is Claude Fable 5 (max) at 35.8%. It covers planning a physical assembly. The board median is 41.2%, so deepseek-v4.1-flash-max is 7.1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

deepseek-v4.1-flash-max lands below the middle on LMArena Text Elo: 1,474, #22 of 28, with 75% of the field scoring higher. gemini-4-argon-high leads at 1,525; the closest result above is claude-opus-4-5-20251101-high-32k at 1,474. It covers head-to-head chat preference (crowd votes). The board median is 1,478, so deepseek-v4.1-flash-max is 4 points below it. Source: LMArena, captured 2026-10-05.

The full table above also lists the other 5 results, including boards with fewer than ten models, where a rank says little.

Strengths and weaknesses

Software engineering: solid, ahead of 67% of the field on average across 4 benchmarks. Its best board here is WebDev Arena (#13 of 67) and its weakest is LMArena Code Elo (#16 of 30). Reasoning, maths and science: solid, ahead of 73% of the field on average across 6 benchmarks. Its best board here is OTIS Mock AIME (#17 of 189) and its weakest is FrontierMath T4 (#30 of 50). General knowledge and chat: uneven, ahead of 46% of the field on average across 4 benchmarks. Its best board here is Mystery Games (#13 of 59) and its weakest is LMArena Text Elo (#22 of 28).

Release history

The DeepSeek model released before it was DeepSeek V4.1 Flash (high) on 2026-09-09, which scores 48.9.

Reading these numbers

The 15 scores on this page come from 3 sources: Epoch AI Benchmarking Hub (12), LMArena (2), ARC Prize official leaderboard (1). 1 of them come from an official leaderboard and 14 from datasets or third-party runs.

Models ranked nearby

More from DeepSeek

Compare it on the live leaderboard → · How the aggregate score works