HELM Lite - WMT 2014 (ru-en): leaderboard
HELM Lite v1.13.0 WMT 2014 Russian-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. 91 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3 Opus (20240229) | 28.46 | #428 |
| 2 | Llama 3.1 405B Instruct | 28.37 | #470 |
| 3 | Nova Pro | 28.14 | #594 |
| 4 | GPT-4o (2024-05-13) | 28.06 | #337 |
| 5 | Claude v1.3 | 28.03 | #983 |
| 6 | Claude 3.5 Sonnet (20240620) | 27.01 | #308 |
| 7 | Claude 2 | 26.79 | #761 |
| 8 | GPT-4o (2024-08-06) | 26.71 | #343 |
| 9 | Llama 3.2 90B Vision Instruct | 26.61 | #621 |
| 10 | Claude 3.5 Sonnet (20241022) | 26.57 | #302 |
| 11 | Llama 3.1 70B Instruct | 26.47 | #567 |
| 12 | GPT-4 Turbo | 26.42 | #417 |
| 13 | Gemini 1.5 Pro (002) | 26.11 | #368 |
| 14 | Llama 3.3 70B Instruct | 26.06 | #541 |
| 15 | Llama 3 70B | 25.9 | #632 |
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-ru-en · How It Works · Data refreshed daily, snapshot 2026-11-05.