ChartGemma: benchmark results
Provider: Other.
Aggregate Rating 1249 ± 51, rank #9265 of 10916 rated models, from 20 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ThermEval - Human Counting (FLIR) | 3.04 | Mean absolute error (people, 0 or more) on Task 3, the numbe | 77.8 |
| ThermEval - Temperature Estimation at 2 ft | 1.2 | Mean absolute error (degrees Celsius, 0 or more) on Task 7, | 77.3 |
| ThermEval - Temperature Estimation at 6 ft | 1.02 | Mean absolute error (degrees Celsius, 0 or more) on Task 7, | 68.2 |
| ThermEval - Temperature Estimation at 10 ft | 1.08 | Mean absolute error (degrees Celsius, 0 or more) on Task 7, | 63.6 |
| PolyChartQA - MLM-Generated Questions | 28.28 | L-Accuracy (%, times 100), zero-shot prompting, on the 2,175 | 57.1 |
| ThermEval - Thermal Reasoning (One Person) | 36 | Accuracy (times 100, 0-100) on Task 5, ranking one person's | 54.3 |
| ThermEval - Thermal Reasoning (Two People) | 46 | Accuracy (times 100, 0-100) on Task 5, whether the chest, fo | 52.2 |
| PolyChartQA (CoT) | 23.51 | H-Accuracy (%, times 100), chain-of-thought prompting, on th | 50 |
| ThermEval - Temperature Estimation (Region) | 2.33 | Mean absolute error (degrees Celsius, 0 or more) on Task 6, | 45.5 |
| ThermEval - Temperature Estimation (Marker) | 3.84 | Mean absolute error (degrees Celsius, 0 or more) on Task 6, | 43.5 |
| PolyChartQA (CoT) - MLM-Generated Questions | 23.36 | L-Accuracy (%, times 100), chain-of-thought prompting, on th | 42.9 |
| PolyChartQA | 19.27 | H-Accuracy (%, times 100), zero-shot prompting, on the 519 h | 37.5 |
Interactive version: theaggregate.ai/model?slug=chartgemma · How It Works · Data refreshed daily, snapshot 2026-11-08.