Instruction
IFEval
Instruction Following Evaluation — precision in following instructions.
9 models published a score
| # | Model | Company | Score |
|---|---|---|---|
| 1 | Nova Pro | Amazon | 92.1 |
| 2 | Claude Opus 4.5 | Anthropic | 92.0 |
| 3 | Command A | Cohere | 90.9 |
| 4 | Gemini 3.1 Pro | Google DeepMind | 90.0 |
| 5 | GPT-5.2 | OpenAI | 89.4 |
| 6 | AFM Server | Apple | 89.1 |
| 7 | Doubao Seed 2.0 Pro | ByteDance | 87.4 |
| 8 | AFM On-Device | Apple | 85.1 |
| 9 | Gemini 3 Pro | Google DeepMind | 85.0 |