LLMs TigerLive LLM benchmarks

DeepSeek V4 Flash 0731 (max): benchmark scores, price and ranking

DeepSeek V4 Flash 0731 (max) from DeepSeek ranks #47 of 741 LLMs on the LLMs Tiger aggregate benchmark score (53.0), based on 14 benchmark results captured up to 2026-10-05.

Organisation
DeepSeek
Type
Open-weight / local model
Open weights
Yes
Released
2026-07-31 (source: epoch)
Cheapest API price
$0.015 in / $1.28 out per million tokens (via openrouter, observed 2026-10-05T21:47Z)
Aggregate rank
#47 of 741 (score 53.0)

Relative to other models, DeepSeek V4 Flash 0731 (max) is strongest on GPQA Diamond (#27 of 274) and weakest on SimpleQA Verified (#44 of 66).

DeepSeek V4 Flash 0731 (max) benchmark results

DeepSeek V4 Flash 0731 (max) benchmark scores, ranked against all models measured on each benchmark
BenchmarkScoreRankSourceCaptured
GPQA Diamond91.0%#27 of 274Epoch AI Benchmarking Hub2026-10-05
DTBench90.9%#22 of 149Epoch AI Benchmarking Hub2026-10-05
OTIS Mock AIME94.4%#28 of 188Epoch AI Benchmarking Hub2026-10-05
WeirdML63.0%#17 of 112Epoch AI Benchmarking Hub2026-10-05
ALE-Bench1,306#12 of 78Epoch AI Benchmarking Hub2026-10-05
Chess Puzzles33.0%#23 of 124Epoch AI Benchmarking Hub2026-10-05
CritPt16.6%#27 of 111Epoch AI Benchmarking Hub2026-10-05
LMCA41.7%#28 of 111Epoch AI Benchmarking Hub2026-10-05
ARC-AGI-261.4%#27 of 89ARC Prize official leaderboard2026-10-05
Mystery Games34.0%#18 of 57Epoch AI Benchmarking Hub2026-10-05
SciCode49.9%#33 of 100Epoch AI Benchmarking Hub2026-10-05
FrontierMath T1-357.5%#31 of 69Epoch AI Benchmarking Hub2026-10-05
FrontierMath T424.4%#33 of 52Epoch AI Benchmarking Hub2026-10-05
SimpleQA Verified33.6%#44 of 66Epoch AI Benchmarking Hub2026-10-05

DeepSeek V4 Flash 0731 (max) is also listed at 5 other reasoning-effort settings; this page shows the highest-effort row, which is the one the leaderboard keeps by default.

Models ranked nearby

  • #44 Codex + GPT-6 Astra (53.1)
  • #45 o3 + gpt-4.1 (53.0)
  • #46 Claude Code + Fable 5.1 (53.0)
  • #48 Qwen3.5-122B-A10B (53.0)
  • #49 Claude Mythos Preview (Early) (53.0)
  • #50 Solar-Open2-250B (52.9)

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