TimeSeriesExamAgent - WeatherBench 2: leaderboard
Time-series reasoning on the WeatherBench 2 weather exam (95 questions) that TimeSeriesExamAgent (2026) generates from real data with LLM-written, LLM-verified and difficulty-filtered question templates; the series is shown as a plot. Multiple-choice accuracy in percent (chance 25); higher is better.
Metric: Accuracy (%, times 100) on the WeatherBench 2 weather exam (95 questions): multiple-choice questions generated by TimeSeriesExamAgent from real data (Claude Sonnet 4 templates verified by gpt-4o and filtered so that gpt-4o does not score below gpt-4o-mini), time series given as plots, default API parameters; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 5 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 54.7 | #178 |
| 2 | Gemini 2.5 Pro | 45.3 | #276 |
| 3 | O3 Mini | 37.9 | #546 |
| 4 | TimeSeriesExamAgent Qwen2.5-VL-Instruct (checkpoint unspecified) | 36.8 | |
| 5 | Gemma 3 27B (IT) | 23.2 | #1324 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=timeseriesexamagent-weatherbench-2 · How It Works · Data refreshed daily, snapshot 2026-11-21.