Atlas

Benchmarks

← All benchmarks

ForecastBench Tournament Market Brier Index

Forecasting · 2024-09-30

Difficulty-adjusted Brier Index on ForecastBench market questions.

Top models (higher is better)

ModelScore
Grok 4.2072.6
Claude Opus 572.3
GPT-5.6 Sol72.2
Opus 4.872.0
GPT-5.6 Sol Pro71.7
Grok 4.571.7
GPT-5.571.6
Opus 4.570.9
Grok 4.670.9
Gemini 3.6 Flash70.8
Gemini 3.7 Flash70.8
Grok 470.8
Gemini 3 Pro Preview70.7
Grok Beta70.7
GPT-4.570.6
Grok 4 Fast70.6
Gemini 3.1 Pro Preview70.5
Sonnet 470.4
Claude 3.7 Sonnet70.4
GPT-570.3
Grok 4.1 Fast70.2
Opus 4.169.9
Gemini 3 Flash Preview69.9
Claude 3.5 Sonnet (Oct 2024)69.8
o3-mini69.7
o369.6
GPT-4 Turbo69.5
Kimi K369.5
Gemini 2.5 Pro69.3
MiniMax M369.3
Qwen3-235B-A22B69.3
GPT-5.269.2
Sonnet 4.569.1
Sonnet 4.669.1
DeepSeek-V369.0
Haiku 4.568.9
QwQ-32B-Preview68.9
DeepSeek-R168.7
Gemini 3.5 Flash68.7
GLM 4.5 Air68.6
Loading Atlas data…