Time Series Reasoning on ECG QA
69.81AccuracyARTIST
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ARTISTmode=+ SFT + RL2026.02 | 69.81 | 52.67 | |
| ARTISTRL=true, SFT=true2026.02 | 69.81 | — | |
| OpenTSLM-4Bmode=+ SFT2026.02 | 69.5 | 41 | |
| TimeMasterRL=true2026.02 | 69.31 | — | |
| VL-Timefew-shot=true2026.02 | 64.36 | — | |
| VL-Time2026.02 | 58.91 | — | |
| ITFormer-4Bmode=+ SFT2026.02 | 57.31 | 49.91 | |
| ARTISTmode=+ SFT2026.02 | 56.31 | 55.68 | |
| ARTISTSFT=true2026.02 | 56.31 | — | |
| Qwen3-14Bstatistics=w/ statistics2026.02 | 54.58 | 51.79 | |
| GPT-5statistics=w/ statistics2026.02 | 53.96 | 49.62 | |
| Base Model2026.02 | 53.54 | — | |
| ChatTS-14Bmode=+ SFT2026.02 | 53.47 | 24.31 | |
| TimeMaster2026.02 | 52.59 | — | |
| GPT-5statistics=w/o statistics2026.02 | 51.98 | 51.78 | |
| Qwen3-8Bstatistics=w/ statistics2026.02 | 51.86 | 49.06 | |
| Qwen3-14Bstatistics=w/o statistics2026.02 | 50.99 | 50.42 | |
| Qwen3-8Bstatistics=w/o statistics2026.02 | 50.62 | 48.46 | |
| Random Guess2026.02 | 50 | 50 | |
| Llama-3 8Bstatistics=w/o statistics2026.02 | 50 | 48.45 | |
| Llama-3 8Bstatistics=w/ statistics2026.02 | 50 | 48.9 | |
| ChatTS-14Bmode=Base Model2026.02 | 48.02 | 32.07 |