Multi-IF — leaderboard
Multi-IF evaluates model capability on instruction following tasks from the linked upstream source with Average as the primary reported metric.
Metric: Average (self-reported). Source: benchmarklist.com. Status: saturation imminent. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O1 Preview | 87.7 |
| 2 | Llama 3.1 405B Instruct | 85.4 |
| 3 | O1 Mini | 85.3 |
| 4 | GPT-4o | 84.3 |
| 5 | Qwen 2.5 72B Instruct | 83.7 |
| 6 | Llama 3.1 70B | 82.6 |
| 7 | Claude 3.5 Sonnet | 81.7 |
| 8 | GPT-4 | 81.5 |
| 9 | Mistral Large 2 (Nov) Instruct (2411) | 80.5 |
| 10 | Claude 3 Sonnet | 78.2 |
| 11 | Gemini 1.5 Pro | 75.8 |
| 12 | Claude 3 Haiku | 72.9 |
| 13 | Gemini 1.5 Flash | 72.5 |
| 14 | Llama 3.1 8B | 68.8 |
Interactive version: theaggregate.ai/benchmark?slug=multi-if · How the rankings work · Data refreshed daily, snapshot 2026-07-22.