MEGA-Bench - Application: Mathematics: leaderboard
The "Application: Mathematics" subtask of MEGA-Bench. TIGER-Lab multimodal evaluation with 500+ tasks across vision-language understanding, generation, and reasoning with 44 models.
Metric: Average Score (%). Source: huggingface.co. Saturation forecast: Around January 2027. 44 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 63.2 | #123 |
| 2 | GLM-4.1V-9B (Thinking) | 50.8 | #472 (GLM-4.1V-9B) |
| 3 | Gemini 2.0 Flash (Preview) | 49.7 | #370 |
| 4 | Claude 3.5 Sonnet (20241022) | 47.6 | #291 |
| 5 | Claude 3.5 Sonnet (20240620) | 47.5 | #299 |
| 6 | GPT-4o | 44 | #343 |
| 7 | Gemma 3 27B (IT) | 42.2 | #523 |
| 8 | Gemini 1.5 Pro (002) | 41.2 | #349 |
| 9 | InternVL3-38B | 41.2 | #405 |
| 10 | InternVL3-78B | 41.2 | #352 |
| 11 | InternVL3-14B | 41 | #505 |
| 12 | Gemini 1.5 Flash (002) | 38.5 | #583 |
| 13 | InternVL2.5-78B | 37.7 | #431 |
| 14 | Qwen 2 VL 72B | 35.5 | #460 |
| 15 | Llama 4 Scout Base | 35.1 | #707 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-application-mathematics · How It Works · Data refreshed daily, snapshot 2026-10-20.