HELM Capabilities - IFEval: leaderboard
HELM Capabilities run of IFEval, measuring strict instruction-following on verifiable formatting constraints.
Metric: IFEval Strict Acc. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 68 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Grok 3 Mini Beta | 95.1 | #296 |
| 2 | Grok 4 | 94.92 | #181 |
| 3 | GPT-5.1 (2025-11-13) | 93.5 | #101 |
| 4 | GPT-5 Nano | 93.19 | #425 |
| 5 | O4 Mini (2025-04-16) | 92.85 | #179 |
| 6 | GPT-5 Mini (2025-08-07) | 92.7 | #172 |
| 7 | Claude Opus 4 (20250514) | 91.77 | #154 |
| 8 | Llama 4 Maverick Instruct FP8 | 90.76 | #414 |
| 9 | GPT-4.1 Mini | 90.36 | #355 |
| 10 | Gemini 2.5 Flash (Preview 04-17) | 89.8 | #237 |
| 11 | Granite 4.0 H Small | 88.97 | #665 |
| 12 | Grok 3 | 88.42 | #304 |
| 13 | Gemini 3 Pro (Preview) | 87.65 | #66 |
| 14 | Mistral Large 2 (Nov) Instruct (2411) | 87.65 | #465 |
| 15 | GPT-5 | 87.49 | #98 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-capabilities-ifeval · How It Works · Data refreshed daily, snapshot 2026-10-20.