List your company

Weather Forecasting Report Generation on WSInstruct (test)

44BLEU-1WeatherSyn-DPO
5.5215.5125.535.49May 8, 2026

Evaluation results

MethodLinks
WeatherSyn-DPOAccess=Open-source, Backbone=Qwen3-VL-8B, Optimization=DPO2026.054432256259635933322930—————————
WeatherSynAccess=Open-source, Backbone=Qwen3-VL-8B2026.054331256055605411111112—————————
WeatherSyn-RFTAccess=Open-source, Backbone=Qwen3-VL-8B, Optimization=RFT2026.054331256259615916151715—————————
Gemini-3 Pro PreviewAccess=Closed-source, Few-shot=2-shot2026.053728236360636024272931—————————
GPT-4.1-MiniAccess=Closed-source, Few-shot=2-shot2026.05201513494949481111—————————
WeatherQAAccess=Open-source, Backbone=Qwen3-VL-8B2026.05191514503651361111—————————
OmniEarthAccess=Open-source, Backbone=Qwen3-VL-8B2026.05171714514051401111—————————
GPT-5-NanoAccess=Closed-source, Few-shot=2-shot2026.05161310504750483322—————————
GPT-5.2 ThinkingAccess=Closed-source, Few-shot=2-shot2026.05121211514952496775—————————
Claude-3.7-SonnetAccess=Closed-source, Few-shot=2-shot2026.0591017535152502111—————————
Gemini-2.5-FlashAccess=Closed-source, Few-shot=zero-shot2026.0571111514451442111—————————
Claude-3.7-Sonnetshot=2-shot2026.05———————————395545315866676953
Gemini-2.5-flashshot=zero-shot2026.05———————————355544335952607651
GPT-4.1-minishot=2-shot2026.05———————————355237296256537149
GPT-5-nanoshot=2-shot2026.05———————————365337326348537850
OmniEarth2026.05———————————355742357256526351
WeatherQA2026.05———————————345549476047496450
WeatherSyn2026.05———————————426263437564676660
WeatherSyn-DPO2026.05———————————446367497764726762
WeatherSyn-RFT2026.05———————————446269477564746662