IFEval: leaderboard, scores and methodology
Verifiable instruction-following tasks with programmatic checks.
Top 10
Bars are scaled to the highest observed score. The leader sets the pace; the spread below it shows how much separation exists in this benchmark.
Leaderboard
| # | Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.5 Pro | Anthropic | 200000 | $20.98 | 89.1 |
| 2 | Claude Opus 4.5 Reasoning | Anthropic | 2000000 | $20.91 | 88.2 |
| 3 | GPT-5.2 Pro | OpenAI | 2000000 | $21.32 | 89.6 |
| 4 | GPT-5.2 Max | OpenAI | 1000000 | $21.39 | 89.2 |
| 5 | GPT-5.2 Reasoning | OpenAI | 1000000 | $20.98 | 87.5 |
| 6 | GPT-5.1 Pro | OpenAI | 200000 | $19.77 | 87.8 |
| 7 | Gemini 3 Pro Max | Google DeepMind | 2000000 | $20.57 | 87.9 |
| 8 | GPT-5.1 Reasoning | OpenAI | 400000 | $20.27 | 88.3 |
| 9 | Claude Opus 4.5 Max | Anthropic | 1000000 | $20.90 | 87.8 |
| 10 | o4 Reasoning | OpenAI | 1000000 | $19.69 | 87.5 |
| 11 | Claude Sonnet 4.5 Pro | Anthropic | 1000000 | $18.73 | 86.7 |
| 12 | GPT-5.2 Vision | OpenAI | 1000000 | $19.03 | 87.4 |
| 13 | Gemini 3 Pro Reasoning | Google DeepMind | 200000 | $19.69 | 87.5 |
| 14 | Gemini 3 Pro | Google DeepMind | 2000000 | $18.24 | 86.1 |
| 15 | Qwen3-Max Max | Alibaba Cloud | 400000 | $18.53 | 85.9 |
| 16 | GPT-5.2 | OpenAI | 2000000 | $19.71 | 88.0 |
| 17 | GPT-5.1 Turbo | OpenAI | 2000000 | $19.19 | 86.1 |
| 18 | Gemini 3 Pro Turbo | Google DeepMind | 400000 | $19.62 | 86.4 |
| 19 | Grok 4 Reasoning | xAI | 2000000 | $19.67 | 87.8 |
| 20 | Qwen3-Max Pro | Alibaba Cloud | 400000 | $18.87 | 86.2 |
| 21 | Claude Sonnet 4.5 Max | Anthropic | 1000000 | $19.90 | 87.2 |
| 22 | Gemini 3 Flash Pro | Google DeepMind | 400000 | $17.71 | 84.8 |
| 23 | o4 | OpenAI | 2000000 | $17.74 | 85.4 |
| 24 | GPT-5.1 Vision | OpenAI | 400000 | $17.76 | 84.8 |
| 25 | Gemini 3 Pro Pro | Google DeepMind | 200000 | $19.80 | 87.3 |
| 26 | Claude Opus 4.5 Vision | Anthropic | 200000 | $17.70 | 85.9 |
| 27 | Grok 3 Max | xAI | 2000000 | $17.66 | 85.9 |
| 28 | o4 Turbo | OpenAI | 2000000 | $19.34 | 86.6 |
| 29 | Claude Opus 4.5 | Anthropic | 400000 | $19.42 | 87.7 |
| 30 | Grok 4 Max | xAI | 2000000 | $19.70 | 86.6 |
| 31 | Qwen3-Max | Alibaba Cloud | 400000 | $17.37 | 84.4 |
| 32 | o3 Max | OpenAI | 2000000 | $17.82 | 84.8 |
| 33 | Claude Opus 4.5 Turbo | Anthropic | 2000000 | $19.57 | 85.7 |
| 34 | GPT-4.1 Pro | OpenAI | 400000 | $17.56 | 85.2 |
| 35 | GPT-5.1 Max | OpenAI | 2000000 | $20.50 | 87.9 |
| 36 | Qwen3-Max Reasoning | Alibaba Cloud | 2000000 | $18.12 | 85.9 |
| 37 | GPT-4.1 Max | OpenAI | 400000 | $17.34 | 84.7 |
| 38 | Qwen3-Max Turbo | Alibaba Cloud | 1000000 | $17.70 | 84.7 |
| 39 | GPT-5.2 Turbo | OpenAI | 1000000 | $20.73 | 87.4 |
| 40 | Gemini 2.5 Pro Pro | Google DeepMind | 400000 | $17.16 | 83.6 |
| 41 | Claude Sonnet 4.5 Reasoning | Anthropic | 400000 | $19.48 | 87.7 |
| 42 | GPT-4.1 Reasoning | OpenAI | 1000000 | $16.97 | 83.8 |
| 43 | Grok 4 | xAI | 200000 | $17.90 | 86.5 |
| 44 | Kimi K2 | Moonshot AI | 128000 | $5.85 | 83.0 |
| 45 | Claude Sonnet 4.5 Turbo | Anthropic | 1000000 | $18.37 | 85.9 |
| 46 | DeepSeek-V3.2 Instruct | DeepSeek | 128000 | $5.91 | 83.7 |
| 47 | Gemini 2.5 Pro Max | Google DeepMind | 200000 | $16.94 | 84.8 |
| 48 | Gemini 3 Flash Max | Google DeepMind | 1000000 | $17.43 | 85.0 |
| 49 | Grok 4 Pro | xAI | 1000000 | $19.00 | 85.4 |
| 50 | Claude Sonnet 4.5 | Anthropic | 400000 | $17.84 | 86.0 |
Methodology
IFEval is included as a instruction following benchmark with a typical score range of 0 to 100. The values in this index are deterministic modelled estimates generated for comparison at scale. They are not official leaderboard results and should not be quoted as vendor benchmarks.
Each model receives a score derived from its capability tier, release generation and benchmark difficulty. Scores are normalised to a 0–100 scale so that models from different providers can be placed side by side. The average shown is the arithmetic mean across all 600 scored models.
Read the leaderboard as a relative map. A five-point gap is meaningful at the top and less meaningful in the middle of the distribution. Pair the score with price and context data before choosing a model.
How to use this leaderboard
- Use the top 10 to shortlist candidates, not to declare a winner.
- Check whether the leader is open weights or proprietary before assuming you can deploy it.
- Compare the score against input and output price on each model page.
- Look at the model's scores on related benchmarks before trusting a single number.
FAQ
What is IFEval?
Verifiable instruction-following tasks with programmatic checks.
Who leads IFEval?
Claude Opus 4.5 Pro from Anthropic leads this index with a score of 96.1, followed by Claude Opus 4.5 Reasoning at 95.9.
What is a good IFEval score?
The top quartile starts around 86.2. The index average is 78.6, so a score above 86.2 is a strong result relative to this dataset.
How are IFEval scores calculated?
The values are modelled estimates on a 0–100 scale, generated from capability tier, release generation and benchmark difficulty. They are not official leaderboard results.
Which models are included?
All 600 models in the index are scored on IFEval, across 40 providers and every model category.