EuroEval English NLU: leaderboard
English-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: NLU Average Score (%). Source: euroeval.com. Saturation forecast: Around April 2027. 341 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash (Preview) (Thinking) | 78.47 | #82 (Gemini 3 Flash (Preview)) |
| 2 | Gemini 3 Flash (Preview) (Non-reasoning) | 77.61 | #82 (Gemini 3 Flash (Preview)) |
| 3 | Gemma 4 31B (IT) | 77.22 | #176 |
| 4 | Claude Sonnet 4.6 | 77.12 | #90 |
| 5 | Gemini 3.1 Flash Lite (Preview) | 76.79 | #194 |
| 6 | Gemini 3 Pro (Preview) | 76.03 | #66 |
| 7 | Grok 4.1 Fast (Reasoning) | 75.92 | #217 (Grok 4.1 Fast) |
| 8 | Gemma 4 26B A4B (IT) | 75.35 | #297 |
| 9 | GPT-5.4 Mini (2026-03-17) (High) | 75.28 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 10 | Llama 3.1 70B | 75.12 | #592 |
| 11 | Qwen 3.5 35B A3B Base | 74.41 | #310 |
| 12 | NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 74.33 | #679 |
| 13 | Llama 3.1 405B | 74.19 | #466 |
| 14 | Qwen 3.5 35B A3B FP8 | 73.94 | #362 |
| 15 | GLM-4.7 Flash | 73.87 | #509 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-english-nlu · How It Works · Data refreshed daily, snapshot 2026-10-20.