HELM SeaHELM - Flores (id-en): leaderboard
HELM SeaHELM Indonesian-to-English machine translation (ChrF++).
Metric: ChrF++. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.7 Sonnet (20250219) | 67.8 | #201 |
| 2 | GPT-4o (2024-11-20) | 67.46 | #381 |
| 3 | Llama 4 Scout Instruct | 67.19 | #539 |
| 4 | GPT-4o Mini (2024-07-18) | 66.52 | #558 |
| 5 | Qwen 2 72B Instruct | 65.8 | #576 |
| 6 | Llama 3.1 8B CPT SEA LIONv3 Instruct | 65.22 | #810 |
| 7 | Qwen 2.5 72B Instruct | 64.08 | #454 |
| 8 | DeepSeek R1 Distill Llama 8B | 56.47 | #1291 |
| 9 | Gemini 2.0 Flash | 54.27 | #342 |
| 10 | Claude 3.5 Haiku (20241022) | 49.4 | #588 |
| 11 | Llama 3.3 70B Instruct | 49.17 | #534 |
| 12 | DeepSeek V3 | 48.78 | #321 |
| 13 | Llama 4 Maverick Instruct FP8 | 48.75 | #414 |
| 14 | Llama 3.2 90B Vision Instruct | 40.04 | #622 |
| 15 | Llama 3.1 70B Instruct | 29.59 | #561 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-seahelm-flores-id-en · How It Works · Data refreshed daily, snapshot 2026-10-20.