# Best LLMs for Translation in 2026

> AI models ranked for translation and multilingual work, from BenchLM's multilingual benchmark category.

Translation quality tracks the multilingual category: benchmarks that test comprehension and generation across languages. The frontier models at the top of this table are effectively tied on major-language pairs; the gaps show up in lower-resource languages, idiom, and domain terminology.

Canonical page: https://benchlm.ai/best/translation

Last updated: September 10, 2026

## Rankings

| Rank | Model | Creator | Type | Context | Score |
|------|-------|---------|------|---------|-------|
| 1 | [Qwen3.7 Max](/models/qwen3-7-max) | Alibaba | Proprietary | 1M | 100 |
| 2 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | Proprietary | 200K | 82.9 |
| 3 | [Qwen3.7 Plus](/models/qwen3-7-plus) | Alibaba | Proprietary | 1M | 78.9 |
| 4 | [Qwen3.6 Plus](/models/qwen3-6-plus) | Alibaba | Proprietary | 1M | 69.7 |
| 5 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | Open Weight | 128K | 69.7 |
| 6 | [GLM-5](/models/glm-5) | Z.AI | Open Weight | 200K | 48.7 |
| 7 | [Nemotron 3 Ultra](/models/nemotron-3-ultra) | NVIDIA | Open Weight | 1M | 47.4 |
| 8 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | Open Weight | 256K | 38.2 |
| 9 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | Open Weight | 262K | 36.8 |
| 10 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | Alibaba | Open Weight | 262K | 36.8 |
| 11 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | Open Weight | 262K | 21.1 |
| 12 | [Qwen3 235B 2507](/models/qwen3-235b-2507) | Alibaba | Open Weight | 128K | 1 |

## Key Takeaways

- Top model: [Qwen3.7 Max](/models/qwen3-7-max) with a score of 100
- Best open-weight option: [Qwen3.5 397B](/models/qwen3-5-397b) at #5
- Models included: 12

## Compare the Leaders

- [Qwen3.7 Max vs Claude Opus 4.5](/compare/claude-opus-4-5-vs-qwen3-7-max)