HELM Lite - WMT 2014 (fr-en): leaderboard
HELM Lite v1.13.0 WMT 2014 French-to-English machine translation (503 sentences evaluated), 1-shot, scored by BLEU-4 on a 0-100 scale; higher is better. One of the five language pairs averaged into the HELM WMT 2014 board.
Metric: BLEU-4 (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 90 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3 Opus (20240229) | 25.23 | #428 |
| 2 | Gemini 1.5 Pro (001) | 25.15 | #465 |
| 3 | Gemini 1.5 Pro (002) | 24.56 | #368 |
| 4 | Llama 3.1 405B Instruct | 24.4 | #470 |
| 5 | Nova Pro | 24.24 | #594 |
| 6 | Gemini 2.0 Flash (Preview) | 24.21 | #376 |
| 7 | Gemini 1.5 Flash (001) | 24.07 | #580 |
| 8 | GPT-4o (2024-05-13) | 24.02 | #337 |
| 9 | Claude 3.5 Sonnet (20241022) | 23.71 | #302 |
| 10 | Llama 3.2 90B Vision Instruct | 23.54 | #621 |
| 11 | Claude 3.5 Sonnet (20240620) | 23.53 | #308 |
| 12 | GPT-4o (2024-08-06) | 23.52 | #343 |
| 13 | Gemma 2 27B (IT) | 23.51 | #652 |
| 14 | Llama 3.1 70B Instruct | 23.49 | #567 |
| 15 | Command-R+ | 23.34 | #869 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-lite-wmt-2014-fr-en · How It Works · Data refreshed daily, snapshot 2026-11-05.