HELM WMT 2014 — leaderboard
Machine translation quality on WMT 2014 benchmark across multiple language pairs, scored by BLEU-4. Part of Stanford HELM.
Metric: BLEU-4 (%). Source: crfm.stanford.edu. Status: saturation imminent. 91 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Palmyra X V3 (72B) | 26.19 |
| 2 | PaLM-2 (Unicorn) | 25.95 |
| 3 | PaLM-2 (Bison) | 24.08 |
| 4 | Claude 3 Opus (20240229) | 23.99 |
| 5 | Palmyra X V2 (33B) | 23.88 |
| 6 | Llama 3.1 405B Instruct | 23.81 |
| 7 | Gemini 1.5 Pro (002) | 23.13 |
| 8 | GPT-4o (2024-05-13) | 23.07 |
| 9 | Claude 3.5 Sonnet (20240620) | 22.9 |
| 10 | Nova Pro | 22.89 |
| 11 | Claude 3.5 Sonnet (20241022) | 22.56 |
| 12 | GPT-4o (2024-08-06) | 22.51 |
| 13 | Gemini 1.5 Flash (001) | 22.51 |
| 14 | Llama 3 70B | 22.46 |
| 15 | Llama 3.2 90B Vision Instruct | 22.41 |
Interactive version: theaggregate.ai/benchmark?slug=helm-wmt-2014 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.