1stTop pick
Researched 25d agoCommand A+
- cESA Avg. Rank (lower is better)
- 26
- Output (from)
- $1.50 / 1M
Best LLM for translation
As of 2026-11-10, LLM Reference's top pick for translation is Command A+ by Cohere (cESA Avg. Rank (lower is better) 26, $1.50/1M output tokens), chosen from models that meet the podium eligibility rules. Second pick: Qwen3.6-27B by Alibaba.
Last refreshed 2026-11-10. Next refresh: weekly.
Models ranked by their WMT26 human-evaluation average rank across language pairs, where lower is better. Preview and retired models, and models without a WMT26 score, are listed below, unranked.
Verdict
Use Command A+ for translation today.
Qwen3.6-27B is the runner-up: 26 vs 35.1 on cESA Avg. Rank (lower is better).
2ndShortlist
Researched 25d agoQwen3.6-27B
- cESA Avg. Rank (lower is better)
- 35.1
- Output (from)
- $2.70 / 1M
3rdShortlist
Researched 16d agoQwen3.5-9B
- cESA Avg. Rank (lower is better)
- 41.5
- Output (from)
- $0.150 / 1M
How we rank
Models with a WMT26 human-evaluation score are ranked by their average rank across language pairs, where lower is better. Preview and retired models, and models without a WMT26 score, are listed below, unranked.
- Eligibility — Chat models with a WMT26 human-evaluation score and a live provider route, plus a curated list of dedicated text translation models and general chat-completion models with at least 128K measured context and at least 7B parameters where size is known. Embedding, ranking, classification, image, video, OCR and audio-only specialists are excluded. Retired models are listed only when they have a WMT26 score, below the ranking, unranked.
- Ranking — WMT26 human-evaluation average rank across language pairs, lower first; self-reported scores do not count. Ties go to the newer release. Preview and retired models with a score are listed below the ranked table with their score and a Preview or Retired badge, unranked.
- Podium eligibility — Preview, restricted, retired and retiring models never fill a podium slot. Preview means preview or invite-only access, or a model that is announced or pre-release. Restricted means restricted, partner, private or otherwise unavailable access. Retired means the model is past its retirement date or every provider route is retired. Retiring means the model, or any one of its provider routes, has a retirement date within the next 60 days. These models can still appear in the table below the podium.
- Pricing column — Lowest listed provider input and output price where a public rate card exists.
| # | Model | Input $/1M | Output $/1M | |
|---|---|---|---|---|
| 1 | Command A+ ReasoningVisionTools cESA Avg. Rank (lower is better): 26 | $0.30 | $1.50 | |
| 2 | Qwen3.6-27B ReasoningVisionTools cESA Avg. Rank (lower is better): 35.1 | $0.32 | $2.70 | |
| 3 | Qwen3.5-9B VisionTools cESA Avg. Rank (lower is better): 41.5 | $0.10 | $0.15 |
Scored but unranked (preview or retired)
| Model | cESA Avg. Rank (lower is better) | Status |
|---|---|---|
| Gemini 3.1 Pro Preview | 4.8 | Preview |
Scored but unranked (not eligible to rank)
| Model | Input $/1M | Output $/1M | |
|---|---|---|---|
| Qwen-MT-Lumen cESA Avg. Rank (lower is better): 6.8 | — | — | |
| GPT-5.5 ReasoningVisionToolsVerify pricing cESA Avg. Rank (lower is better): 8.6 | $5.00 | $30.00 | |
| DeepSeek V4 Pro ReasoningToolsVerify pricing cESA Avg. Rank (lower is better): 17.5 | $0.36 | $0.73 | |
| Gemma 4 31B IT VisionToolsVerify pricing cESA Avg. Rank (lower is better): 22 | Free | Free | |
| Mistral Medium 3.5 ReasoningVisionToolsVerify pricing cESA Avg. Rank (lower is better): 29.3 | $1.50 | $7.50 | |
| gpt-oss-120b ToolsVerify pricing cESA Avg. Rank (lower is better): 32.1 | Free | Free | |
| gpt-oss-20b ToolsVerify pricing cESA Avg. Rank (lower is better): 39.8 | Free | Free |
Dedicated translation models (no WMT26 score)
Unranked. No WMT26 human-evaluation score.
| Model | Input $/1M | Output $/1M | |
|---|---|---|---|
| Hy-MT2 30B-A3B cESA Avg. Rank (lower is better): — | $0.07 | $0.29 | |
| Hy-MT2 7B cESA Avg. Rank (lower is better): — | $0.07 | $0.29 | |
| Qwen-MT-Plus Verify pricing cESA Avg. Rank (lower is better): — | $0.25 | $0.75 | |
| Qwen-MT-Flash Verify pricing cESA Avg. Rank (lower is better): — | $0.16 | $0.49 | |
| Qwen-MT-Lite Verify pricing cESA Avg. Rank (lower is better): — | $0.12 | $0.36 | |
| Command A Translate (08-2025) cESA Avg. Rank (lower is better): — | — | — |
General chat models with long context
Unranked. No WMT26 human-evaluation score. Newest release first.
| Model | Input $/1M | Output $/1M | |
|---|---|---|---|
| K-EXAONE 2.5 ReasoningTools cESA Avg. Rank (lower is better): — | — | — | |
| Claude Fable 5.2 ReasoningVisionTools cESA Avg. Rank (lower is better): — | $8.00 | $40.00 | |
| Claude Mythos 5.2 RestrictedReasoningVisionTools cESA Avg. Rank (lower is better): — | — | — | |
| Muse Ember ReasoningVision cESA Avg. Rank (lower is better): — | $0.10 | $0.50 | |
| ERNIE 5.5 Flash Base cESA Avg. Rank (lower is better): — | — | — | |
| Inkling 2 Small ReasoningVisionTools cESA Avg. Rank (lower is better): — | Free | Free | |
| Ling-3.2-flash VisionTools cESA Avg. Rank (lower is better): — | — | — | |
| Ripple-65B Preview PreviewTools cESA Avg. Rank (lower is better): — | $0.20 | $0.60 | |
| Particle-1 ReasoningVisionTools cESA Avg. Rank (lower is better): — | $0.25 | $1.00 | |
| Gemini 4 Carbon ReasoningVisionTools cESA Avg. Rank (lower is better): — | $2.00 | $10.00 | |
| GPT-6.1 Terra ReasoningVisionTools cESA Avg. Rank (lower is better): — | $1.00 | $5.00 | |
| Nanbeige4.3-33B ReasoningTools cESA Avg. Rank (lower is better): — | — | — | |
| Kimi K3.1 Pro ReasoningVisionTools cESA Avg. Rank (lower is better): — | $2.00 | $10.00 | |
| Kimi K3.1 Pro Highspeed ReasoningVisionTools cESA Avg. Rank (lower is better): — | $4.00 | $20.00 |
Compare Top Picks
Side-by-side comparison of the top picks by price, benchmark, and API access.
Browse Other Categories
Best LLMs for Code GenerationBest LLMs for RAGBest LLMs for Agents vs Agent Products (2026)Best LLMs for ClassificationBest Open Source LLMsBest Multimodal / Vision LLMsBest AI Image Models in 2026Best AI Video Models in 2026Best LLMs for Reasoning & MathBest Small Language Models (SLMs)Best LLMs for JSON / Tool useCheapest LLM APIs You Can Call Right NowBest Long Context LLMsBest Mainstream LLM APIs, RankedBest LLMs for EnterpriseBest Free LLMs You Can Use Right NowBest LLMs for WritingBest LLMs for MarketingBest LLMs for Customer Support
Frequently asked questions
What is the best LLM for translation?
LLM Reference's top pick for translation as of 2026-11-10 is Command A+ by Cohere, with cESA Avg. Rank (lower is better) 26. Picks come from models that meet the podium eligibility rules, so the top pick may not lead the full table. Second pick: Qwen3.6-27B by Alibaba (cESA Avg. Rank (lower is better) 35.1).
Which ranked LLM for translation has the lowest output price?
Among the ranked models shown on this page, Qwen3.5-9B by Alibaba has the lowest listed output price: $0.150/1M output tokens, at rank 3.
What is the best open-weight LLM for translation?
Command A+ by Cohere is the highest-ranked open-weight model shown on this page (rank 1, cESA Avg. Rank (lower is better) 26). License: Apache 2.0.
How is the translation ranking calculated?
Models with a WMT26 human-evaluation score are ranked by their average rank across language pairs, where lower is better. Preview and retired models, and models without a WMT26 score, are listed below, unranked. Preview, restricted, retired and retiring models never fill a podium slot. Preview means preview or invite-only access, or a model that is announced or pre-release. Restricted means restricted, partner, private or otherwise unavailable access. Retired means the model is past its retirement date or every provider route is retired. Retiring means the model, or any one of its provider routes, has a retirement date within the next 60 days. These models can still appear in the table below the podium.
How often is this page updated?
Weekly. Last refreshed 2026-11-10.