MIBench-mini (Multimodal Interaction) - Synergy: leaderboard
Metric: Accuracy (%) on the synergy samples (both modalities are needed) on MIBench-mini, the paper's subset on which the closed-source models were also evaluated (multiple-choice image-text tasks; vision- and text-centric samples average several context variations); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 29 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 81 | #64 |
| 2 | Gemini 2.5 Pro | 76 | #145 |
| 3 | Claude Sonnet 4.5 | 72 | #138 |
| 4 | O3 (2025-04-16) | 72 | #117 |
| 5 | GPT-5.1 (2025-11-13) | 71 | #97 |
| 6 | GPT-5.2 | 70 | #105 |
| 7 | Qwen 3 VL 235B A22B Instruct | 60 | #264 |
| 8 | Qwen 2.5 VL 72B Instruct | 59 | #364 |
| 9 | GPT-4o (2024-11-20) | 57 | #369 |
| 10 | Qwen 3 VL 8B Instruct | 52 | #401 |
| 11 | Qwen 2.5 VL 32B Instruct | 50 | #443 |
| 12 | Qwen 2.5 VL 7B Instruct | 49 | #643 |
| 13 | InternVL3-8B | 45 | #606 |
| 14 | InternVL2-8B | 45 | #826 |
| 15 | Qwen 2 VL 7B Instruct | 43 | #816 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=mibench-mini-multimodal-interaction-synergy · How It Works · Data refreshed daily, snapshot 2026-10-11.