Forecasting on Spatio-Temporal Synthetic Dataset 1.0 (test)
63.36MAEClaude-4.5-Sonnet
Evaluation results
| Method | Links | |
|---|---|---|
| Claude-4.5-SonnetCategory=Proprietary Models, Prompting=Image, Tokens (M)=0.48, Est. Cost ($)=17.482026.01 | 63.36 | |
| Claude-4.5-SonnetCategory=Proprietary Models, Prompting=Text, Tokens (M)=0.74, Est. Cost ($)=45.802026.01 | 63.74 | |
| GPT-5.2Category=Proprietary Models, Prompting=Text, Tokens (M)=0.49, Est. Cost ($)=22.482026.01 | 63.99 | |
| GPT-5.2Category=Proprietary Models, Prompting=Image, Tokens (M)=0.31, Est. Cost ($)=6.692026.01 | 64.7 | |
| STReasoner-8B (Ours)Category=Spatio-Temporal RMs, Prompting=TS Encoder, Tokens (M)=0.08, Est. Cost ($)=0.272026.01 | 65.59 | |
| Qwen3-8B - SFT+S-GRPOCategory=Open-Source Models, Prompting=Text, Tokens (M)=0.72, Est. Cost ($)=3.852026.01 | 66.35 | |
| Qwen3-8B - SFTCategory=Open-Source Models, Prompting=Text, Tokens (M)=0.72, Est. Cost ($)=3.852026.01 | 66.49 | |
| Qwen3-VL-8B-Instruct - SFT+S-GRPOCategory=Open-Source Models, Prompting=Image, Tokens (M)=0.26, Est. Cost ($)=0.662026.01 | 67.29 | |
| Time-R1-7BCategory=Time Series RMs, Prompting=Text, Tokens (M)=0.72, Est. Cost ($)=3.852026.01 | 68.15 | |
| Qwen3-VL-8B-Instruct - SFTCategory=Open-Source Models, Prompting=Image, Tokens (M)=0.26, Est. Cost ($)=0.662026.01 | 68.53 | |
| Qwen3-VL-8B-InstructCategory=Open-Source Models, Prompting=Image, Tokens (M)=0.26, Est. Cost ($)=0.662026.01 | 74.21 | |
| Time-MQA-7BCategory=Time Series LMs, Prompting=Text, Tokens (M)=0.72, Est. Cost ($)=3.852026.01 | 84.7 | |
| ChatTS-8BCategory=Time Series LMs, Prompting=TS Encoder, Tokens (M)=0.08, Est. Cost ($)=0.272026.01 | 85.14 | |
| Qwen3-8BCategory=Open-Source Models, Prompting=Text, Tokens (M)=0.72, Est. Cost ($)=3.852026.01 | 94.03 |