Atlas

Benchmarks

← All benchmarks

ForecastBench Baseline Overall Brier Index

Forecasting · 2024-09-30

Equal-weighted Brier Index for the ForecastBench baseline.

Top models (higher is better)

ModelScore
o361.5
Kimi K361.0
MiniMax M360.9
Opus 4.160.6
Sonnet 4.660.6
GPT-4.560.6
O4 Mini60.6
Claude 3.7 Sonnet60.6
Gemini 2.5 Pro Preview 03-2560.5
Grok 4.2060.5
GPT-5.560.4
GPT-4.160.3
Opus 460.1
Sonnet 4.560.1
GLM-560.0
GPT-560.0
Gemini 2.5 Flash Preview 04-1759.9
Gemini 3 Pro Preview59.9
Grok 4.359.9
Claude 3.5 Sonnet (Oct 2024)59.7
Grok 4.1 Fast59.7
Grok Beta59.7
Grok 459.5
Opus 4.559.4
Opus 4.759.4
Sonnet 559.4
Haiku 4.559.3
Sonnet 459.3
Gemini 2.5 Pro59.2
Grok 4 Fast59.2
Opus 4.859.1
GPT-5 Mini59.1
Gemini 3 Flash Preview59.0
DeepSeek-R158.9
GPT-5.258.9
Opus 4.658.8
Gemini 3.1 Pro Preview58.8
Gemini 3.5 Flash58.8
Qwen3-235B-A22B58.8
Llama 3.1 405B Instruct58.7
Loading Atlas data…