ParseBench - Charts: leaderboard
Metric: ChartDataPointMatch Score. Source: parsebench.ai. Saturation forecast: Around February 2028. 78 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5.1 | 67.06 |
| 2 | GPT-5.5 (Medium) | 65.53 |
| 3 | GPT-5.4 (Non-reasoning) | 65.22 |
| 4 | Gemini 3 Flash (Minimal) | 64.83 |
| 5 | Gemini 3 Flash (High) | 64.79 |
| 6 | Claude Opus 5.5 | 64.12 |
| 7 | GPT-5.6 Terra (Non-reasoning) | 63.75 |
| 8 | Claude Sonnet 5 | 60.53 |
| 9 | GPT-5.6 Sol (Non-reasoning) | 59.57 |
| 10 | GPT-5.5 (Non-reasoning) | 59.11 |
| 11 | Claude Opus 4.7 | 55.84 |
| 12 | GLM-5.3 Flash | 54.96 |
| 13 | Claude Fable 5 | 52.21 |
| 14 | DeepSeek V4.1 Flash (High) | 51.11 |
| 15 | Claude Opus 4.8 | 49.75 |
Interactive version: theaggregate.ai/benchmark?slug=parsebench-charts · How It Works · Data refreshed daily, snapshot 2026-09-24.