Global Pattern Perception on MMTS-Bench
97.5AccuracyLLATISA
Evaluation results
| Method | Links | |
|---|---|---|
| LLATISAModality=Vision (plot + num)2026.04 | 97.5 | |
| GPT-4oModality=Vision (plot)2026.04 | 96.7 | |
| GPT-4oModality=Vision + Text (w/ index)2026.04 | 96.7 | |
| GPT-4oModality=Text (w/o index)2026.04 | 95 | |
| GPT-4oModality=Vision (plot + num)2026.04 | 95 | |
| GPT-4oModality=Vision + Text (w/o index)2026.04 | 94.2 | |
| Qwen3-VL-8BModality=Vision + Text (w/ index)2026.04 | 93.3 | |
| Qwen3-VL-8BModality=Vision + Text (w/o index)2026.04 | 92.5 | |
| Qwen3-VL-8BModality=Vision (plot + num)2026.04 | 86.7 | |
| Qwen3-VL-8BModality=Vision (plot)2026.04 | 85.8 | |
| ChatTSModality=Time Series2026.04 | 80 | |
| Qwen3-8BModality=Text (w/o index)2026.04 | 77.5 | |
| Time-R1Modality=Text (w/o index)2026.04 | 46.7 | |
| Time-MQAModality=Text (w/o index)2026.04 | 28.3 | |
| OpenTSLMModality=Time Series2026.04 | 15 |