GIM: leaderboard
Grounded Integration Measure from Meta FAIR: 820 multimodal and text-grounded problems requiring integrated reasoning across quantitative, spatial, language, world-knowledge, and document tasks.
Metric: IRT ability (theta). Source: arxiv.org. Status: saturation imminent. 46 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 Pro (xHigh) | 2.16 |
| 2 | Gemini 3.1 Pro (Preview) (High) | 2.04 |
| 3 | Gemini 3.1 Pro (Preview) (Medium) | 1.85 |
| 4 | GPT-5.4 (xHigh) | 1.64 |
| 5 | GPT-5.4 (High) | 1.59 |
| 6 | Claude Opus 4.7 (High) | 1.53 |
| 7 | Claude Opus 4.6 (High) | 1.48 |
| 8 | Gemini 3.1 Pro (Preview) (Low) | 1.45 |
| 9 | GPT-5.4 (Medium) | 1.39 |
| 10 | Claude Sonnet 4.6 (High) | 1.12 |
| 11 | GPT-5 (High) | 0.9 |
| 12 | Claude Sonnet 4.6 (Medium) | 0.84 |
| 13 | GPT-5.4 (Low) | 0.82 |
| 14 | GPT-5 (Medium) | 0.59 |
| 15 | Gemini 3.1 Flash Lite (High) | 0.1 |
Interactive version: theaggregate.ai/benchmark?slug=gim · How It Works · Data refreshed daily, snapshot 2026-09-05.