List your company

Time Series Reasoning on WeatherBench 2

54.7AccuracyGPT-5
21.9430.44538.9547.455Apr 11, 2026

Evaluation results

MethodLinks
GPT-52026.0454.7
Gemini-2.5-Pro2026.0445.3
gpt-4o2026.0438.9
o3-mini2026.0437.9
Qwen2.5-VL-Instruct2026.0436.8
random guess2026.0425
Gemma-3-27b-it2026.0423.2