HELM · IFEval
The Frontier
Best score over time · one chart, every benchmark
Distribution
Where models cluster
Correlated benchmarks
Pearson r · original research
Benchmarks that track with HELM · IFEval
Pearson correlation across models scored on both benchmarks. Closer to 1 = strongly predictive.
Full rankings
34 models tested · sorted by score
| # | Model | Score |
|---|---|---|
| 1 |  Grok 3 Mini Beta | 95.1 |
| 2 | Grok 4 | 94.9 |
| 3 |  GPT-5.1 | 93.5 |
| 4 | GPT-5 Nano | 93.2 |
| 5 | o4 Mini | 92.9 |
| 6 | GPT-5 Mini | 92.7 |
| 7 | GPT-4.1 Mini | 90.4 |
| 8 |  Gemini 2.5 Flash | 89.8 |
| 9 | Grok 3 Beta | 88.4 |
| 10 | Gemini 3 Pro | 87.6 |
| 11 |  Mistral Large 2411 | 87.6 |
| 12 | GPT-5 Chat | 87.5 |
| 13 | o3 | 86.9 |
| 14 |  Claude 3.5 Sonnet | 85.6 |
| 15 |  Kimi K2 0711 | 85.0 |
| 16 | GPT-4.1 Nano | 84.3 |
| 17 | Gemini 2.0 Flash | 84.1 |
| 18 | Gemini 2.5 Pro | 84.0 |
| 19 | GPT-4.1 | 83.8 |
| 20 | Gemini 1.5 Pro (Feb 2024) | 83.7 |
| 21 | gpt-oss-120b | 83.6 |
| 22 | Claude 3.7 Sonnet | 83.4 |
| 23 |  DeepSeek V3 | 83.2 |
| 24 | Gemini 1.5 Flash (May 2024) | 83.1 |
| 25 | Gemini 2.0 Flash Lite | 82.4 |
| 26 |  Palmyra X5 | 82.3 |
| 27 | GPT-4o (2024-11-20) | 81.7 |
| 28 | Gemini 2.5 Flash Lite | 81.0 |
| 29 |  Qwen3 Next 80B A3B Thinking | 81.0 |
| 30 | Claude 3.5 Haiku | 79.2 |
| 31 | R1 0528 | 78.4 |
| 32 | GPT-4o-mini (2024-07-18) | 78.2 |
| 33 | Mistral Small 3.1 24B | 75.0 |
| 34 | gpt-oss-20b | 73.2 |
Frequently asked
Pulled from the HELM · IFEval dataset · updated daily
What does HELM · IFEval measure?
HELM · IFEval is a knowledge benchmark in the BenchGecko catalog. 34 AI models have been tested on it. Scores range from 73.2 to 95.1 out of 100.
Which model leads on HELM · IFEval?
Grok 3 Mini Beta from xAI leads HELM · IFEval with a score of 95.1. The median score across 34 tested models is 84.0.
Is HELM · IFEval saturated?
Yes · the top model on HELM · IFEval has reached 95.1 out of 100, within 5% of the theoretical ceiling. This benchmark is approaching saturation and may be replaced by a harder successor.
Does HELM · IFEval predict performance on other benchmarks?
Yes · HELM · IFEval scores correlate 0.93 with Cybench across 5 shared models. Models that do well on HELM · IFEval tend to do well on Cybench.
How often is HELM · IFEval data refreshed?
BenchGecko pulls updates daily. New model scores on HELM · IFEval appear as soon as they are published by Epoch AI or the model provider.
More knowledge benchmarks
Same category · related evaluations