DeepSeek V4 Pro (0813): benchmark results

Provider: DeepSeek. Released 2026-08-13. Access: Open.

Unified ELO 1711 ± 1, rank #42 of 1392 rated models, from 81 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LLM Stats (Humanity's Last Exam (with tools, text-only))60Score (%)100
LLM Stats (NL2Repo)61.5Score (%)100
LLM Stats Score52.49LLM Stats Score (conservative rating)97.4
AI Chess Leaderboard (Continuation)1502Elo96.2
BenchmarkList ECI148.12Capability Index (ECI)94.9
OpenRouter Tau2-Bench Airline78Accuracy (%)94.1
LLM Stats (Toolathlon)74.1Score (%)93.8
AI Chess Leaderboard (Reasoning)1484Elo92.3
NL2Repo61.1Score (self-reported)89.7
LiveBench Consecutive Events90.23Score88.7
Gert Labs Rankings59GScore (%)87.9
LiveBench Connections100Score87.7

Interactive version: theaggregate.ai/model?slug=deepseek-v4-pro-0813 · How It Works · Data refreshed daily, snapshot 2026-09-05.