whatmodel.sh / Command A Translate (08-2025)
Command A Translate (08-2025)
↗ share
+ contribute
Command A Translate (08-2025) is an open-weights language model from Cohere, released August 2025. Its published size is 111B parameters and a 8k-token context window. Licence: cc-by-nc-4.0.
Catalogue identity notes
Cohere API id command-a-translate-08-2025 ('Command A Translate is Cohere's state of the art machine translation model ... 23 languages'); weights CohereLabs/command-a-translate-08-2025 @15bf49e (created 2025-08-27, cc-by-nc-4.0, gated). Paper 2609.13916v1 Table 1 row 'Command A Translate (2025)' (111B dense) is this checkpoint, Cohere's only Command A Translate. Only third-party GGUF conversions exist, so local is null.
Model information sources
Release date ↗ · Weights licence ↗
Paper Table 1 '111B (111B)' (Cohere's own table); HF safetensors metadata 111,057,580,032.
$ how it compares
The LLMs ranked either side of Command A Translate (08-2025), out of 828 with a score, then more from Cohere.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 390 |  GPT-4 (0613) | 51±1 | api $37.50/M | ||
| 391 |  OpenCoder-8B-Instruct | 51±1 | local | ||
| 392 | Command A Translate (08-2025) | 51±5 | api price unknown | ||
| 393 |  Llama 3.1 Nemotron 51B Instruct | 51±2 | local | ||
| 394 |  Ministral 3 3B Reasoning 2512 | 51±1 | local | ||
| more from Cohere | |||||
| 76 | North Small Translate 1.0 | 71±6 | api price unknown | ||
| 175 | Command A+ (05-2026) | 64±1 | api price unknown +local | ||
| 277 | Command A (03-2025) | 57±1 | api $4.38/M +local | ||
$ what it's good at
other tests3 not in a task · 90% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| WMT25 Terminology Translation (term success rate) lab | 56.9% ✎ | ▃·▂··▂·▄▂·█▆ beats 0/12 | 55% |
| WMT24++ lab | 76.9% ✎ | █·▃·▅▅▃···▃▃ beats 2/10 | 19% |
| WMT26 General MT (GEMBA-ESA, LLM judge) lab | 70.0% ✎ | ▄·▄▄▄·▄██▄█▄ beats 2/12 | 16% |
$ where to run it
Through an API
api · price unknown · 8k context · docs.cohere.com ↗
Weights and documentation: huggingface.co/CohereLabs ↗
$ how the score was worked out · Cohere — North Small Translat… 90%
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 Cohere — North Small Translate (arXiv:2609.13916v1) · lab-reported | 3 | 90% |
| prior · pull toward the middle | 10% | |
| weighted average | = 51.1 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json