HELM Lite - WMT 2014 (hi-en): leaderboard
HELM Lite v1.13.0 WMT 2014 Hindi-to-English machine translation (832 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 | Llama 3.1 405B Instruct | 20.02 | #470 |
| 2 | Gemini 2.0 Flash (Preview) | 19.83 | #376 |
| 3 | Gemini 1.5 Pro (002) | 19.23 | #368 |
| 4 | Claude 3 Opus (20240229) | 18.78 | #428 |
| 5 | Gemini 1.5 Flash (001) | 18.61 | #580 |
| 6 | Nova Pro | 18.4 | #594 |
| 7 | Llama 3 70B | 18.34 | #632 |
| 8 | Llama 3.1 70B Instruct | 18.33 | #567 |
| 9 | Llama 3.2 90B Vision Instruct | 18.21 | #621 |
| 10 | Claude 3.5 Sonnet (20240620) | 18.14 | #308 |
| 11 | Llama 3.3 70B Instruct | 18.03 | #541 |
| 12 | GPT-4o (2024-08-06) | 17.98 | #343 |
| 13 | Gemini 1.5 Flash (002) | 17.86 | #587 |
| 14 | GPT-4o (2024-05-13) | 17.57 | #337 |
| 15 | Claude 3.5 Sonnet (20241022) | 17.42 | #302 |
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-hi-en · How It Works · Data refreshed daily, snapshot 2026-11-05.