CohereLanguage model

Command A Translate

Command A Translate has no overall index value yet because there are too few independent benchmark results.

PopularityRank 707 of 785

20out of 100same as last month

based on search, usage and attention

Summary

Command A Translate is a Command A specialist built for machine translation. Cohere names 23 languages and positions the checkpoint around translation work rather than general agent conversations.

Cohere divides the stated 16K window into 8,000 input tokens and 8,000 output tokens. The 111-billion-parameter checkpoint supports private deployments and, according to Cohere, can run on one or two A100 or H100 GPUs. Command A remains the general family model for RAG and tool use.

All benchmark results

Every published result with its source, test setup and comparison models.

Show all 80 published benchmark results

Measurements without matching peer values

These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.

WMT25 prompt-injection failure rate, Arabisch (Show measurement, test conditions, and source)

Source value: 0
Score: 0
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; ar
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Tschechisch (Show measurement, test conditions, and source)

Source value: 0
Score: 0
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; cs
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Deutsch (Show measurement, test conditions, and source)

Source value: 0.1
Score: 0.1
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; de
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Griechisch (Show measurement, test conditions, and source)

Source value: 0.1
Score: 0.1
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; el
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Spanisch (Show measurement, test conditions, and source)

Source value: 0
Score: 0
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; es
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Persisch (Show measurement, test conditions, and source)

Source value: 0.2
Score: 0.2
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; fa
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Französisch (Show measurement, test conditions, and source)

Source value: 0.1
Score: 0.1
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; fr
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Hebräisch (Show measurement, test conditions, and source)

Source value: 0
Score: 0
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; he
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Hindi (Show measurement, test conditions, and source)

Source value: 0.2
Score: 0.2
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; hi
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Indonesisch (Show measurement, test conditions, and source)

Source value: 0.2
Score: 0.2
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; id
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Italienisch (Show measurement, test conditions, and source)

Source value: 0.1
Score: 0.1
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; it
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

WMT25 prompt-injection failure rate, Koreanisch (Show measurement, test conditions, and source)

Source value: 2.2
Score: 2.2
Metric: failed translation rate under the terminal-question-mark heuristic
Unit: %
Benchmark version: WMT25 Table 6; ko
Category: translation
Direction: Lower is better
Source type: vendor-reported
Evaluator: WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6
Status: active
Retrieved at: 2026-10-18
Methodology: Vendor-reported result. Exact API model ID command-a-translate-08-2025 is confirmed in Cohere's official model documentation. WMT25 Table 6 evaluates the Command A Translate model, not the two-model CommandA-WMT submission, on adversarial machine-translation prompt injection using a terminal-question-mark heuristic. The paper says Section 3.6 used 817 English questions but Appendix C says 816; it does not state the per-language denominator or all generation settings.
Context: This is a vendor-reported result in a WMT proceedings paper authored by Cohere researchers, not an independent replication. The score is the percentage of failed translations, so lower is better. The source disagrees on whether the English question set contains 816 or 817 items, publishes no per-language denominator, and reports no confidence interval.

Source WMT 2025 paper by Cohere authors, Command-A-Translate, Table 6

Page 1 of 7

Popularity compared

How often the model is searched for, used and mentioned, compared with every model of its category.

Popularity index

The 100 most popular models of this category, 0 to 100, higher is more popular

20 of 101 models

GPT-6 Astra83

Claude Haiku 5.577

GPT-6 Sol73

GPT-573

DeepSeek V4.1 Flash 101769

Kimi K367

MiniCPM-V 4.7 35B-A3B63

GPT-6.1 Sol63

Claude Opus 563

Qwen 3.8 Flash62

Command A Translate20

DeepSeek V4.1 Flash 1017

MiniCPM-V 4.7 35B-A3B

GPT-6.1 Sol

Claude Opus 5

Qwen 3.8 Flash

Command A Translate

As of November 2026 · Source: gradually.ai popularity index

Over time

Weekly values from October 18, 2026 to November 7, 2026

Command A TranslateAya Vision 32B 20Codestral-22B-v0.1 19DBRX 19Llama 3.1 70B 19

Method changes in the timeline, a jump at these points is no market event:

1November 1, 2026: Value only with search or usage data, 30-day attention

What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.

Similar models

More models of the same kind as Command A Translate, chosen by model line, developer, size, price and release date.

All specifications

Architecture

Parameters: 111 billion; Source: Parameters
Knowledge cutoff: June 2024; Source: Knowledge cutoff

Input and output

Inputs: Text; Source: Inputs
Outputs: Text; Source: Outputs
Tool calling: Yes; Source: Tool calling
Combined input and output window: 16K tokens; Source: Combined input and output window
Maximum output: 8,000 tokens; Source: Maximum output

Watermarking and provenance

AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.

AI watermarking: Unknown; no source
Watermarking technique: Not known; no source

Licence and access

Weights: Open weights; Source: Weights
Access: API; Source: Access
License: CC BY-NC-4.0; Source: License

Further details

Maximum input: 8,000 tokens; Source: Maximum input

Frequently asked questions about Command A Translate

What is the knowledge cutoff of Command A Translate?

June 2024, according to Cohere. That is 14 months before release.

Stated by Cohere

JunSepDecMarJunKnowledge cutoffJune 2024ReleasedAug 28, 2025

Can I run Command A Translate locally?

Yes, the weights are open. With 111B parameters it needs a lot of memory, more a server than a laptop.

Stated by Cohere

Open weights: yes
License: CC BY-NC-4.0
Parameters: 111B
active per token: n/a

Sources and data date

Show all 17 sources

Benchmark sources at a glance

Which independent evaluators tested this model. This is coverage, not a quality ranking.

No active result from the reviewed external evaluators can currently be attributed to this model. Other source evidence remains available below.

Without vendor claims and historical or disputed results; those stay labelled in the single values.

Type	Evidence and data date
Input	Documented modalities. Retrieved October 27, 2026. Source: models.dev catalog · models.dev catalog (retrieved October 27, 2026)
Output	Documented modalities. Retrieved October 27, 2026. Source: models.dev catalog · models.dev catalog (retrieved October 27, 2026)
Research status	Research date October 15, 2026. 7 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Not published	decoding-details Note The standard long-context row is identified, but not all inference and prompt details are specified; paragraph-level values remain separate and excluded.
Additional source	Hugging Face · Cohere Command A Translate model card · Editorial description reviewed October 2, 2026
Additional source	Cohere Command A Translate input and output limits (retrieved October 3, 2026) · Cohere Command A Translate documentation · Editorial description reviewed October 2, 2026
Additional source	WMT 2025 paper: Command A Translate long-context evaluation (retrieved October 15, 2026)
Additional source	WMT 2025 Proceedings, Command A Translate paper, Table 2 (retrieved October 17, 2026; October 18, 2026)
Additional source	Alexandria: Context-Aware Machine Translation for Arabic Dialects (retrieved October 17, 2026)
Additional source	Current state of LLMs for Arabic dialectal machine translation (retrieved October 17, 2026)
Additional source	Cohere authors, WMT 2025 paper (retrieved October 15, 2026; October 22, 2026)
Additional source	Model metadata (retrieved October 15, 2026)
Additional source	Model metadata (retrieved October 18, 2026)
Additional source	Cohere current model overview · Editorial description reviewed October 2, 2026
Additional source	OpenRouter: Popularity index (derived), CC BY 4.0
Additional source	Vercel AI Gateway: Popularity index (derived), CC BY 4.0
Additional source	Ollama: cumulative downloads for the popularity index
Additional source	Replicate: cumulative runs for the popularity index
Additional source	Google Trending Now: limited feed selection for Germany and the US
Additional source	Creative Commons Attribution 4.0