OpenVLM MMT-Bench - Counting By Reasoning: leaderboard
The "Counting By Reasoning" subtask of OpenVLM MMT-Bench. OpenCompass OpenVLM evaluation of MMT-Bench: multimodal tasks covering visual recognition, localization, OCR, counting, relation reasoning, visual math, and hallucination detection.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 207 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Sonnet | 95 | #350 |
| 2 | GPT-4.1 (2025-04-14) | 90 | #250 |
| 3 | Claude 3.7 Sonnet | 90 | #249 |
| 4 | Qwen 2 VL 72B | 90 | #460 |
| 5 | Pixtral-12B | 90 | #808 |
| 6 | Grok 2 (1212) | 90 | #496 |
| 7 | Llama 3.2 90B Vision Instruct | 90 | #622 |
| 8 | MiniCPM-V-2.6 | 85 | #836 |
| 9 | Claude 3 Sonnet | 85 | #775 |
| 10 | InternVL3-8B | 80 | #620 |
| 11 | InternVL3-38B | 80 | #405 |
| 12 | Gemini 1.5 Pro | 80 | #445 |
| 13 | Qwen 2.5 VL 7B | 80 | #643 |
| 14 | Llama 3.2 11B Instruct | 80 | #1126 |
| 15 | Gemini 1.5 Flash | 80 | #618 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-counting-by-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-20.