Command A
Cohere · · 367 days after Command R
Verified
3 of 3 published values traced to a source3 primary sources, 3 archived
Why this status
Every value this record asserts — date (Mar 13, 2025), context window (256K) and parameter count (111B) — was found in an archived primary source, checked 2026-08-10. Modalities recorded as text-to-text. The primary source states the model is text-only. Capabilities evidenced from the primary sources: agentic: “com/blog"> < Back to blog Introducing Command A: Max performance, minimal compute Cohere Team Mar 13, 2025 Command A is on par or better than GPT-4o and DeepSeek-V3 ac” | multilingual: “Compared to other leading proprietary and open-weights models Command A delivers maximum performance with minimum hardware costs, excelling on business-critical agentic a” | tool_use: “Tool Use Capabilities: Command A has been specifically trained with conversational tool use capabilities.”.
What these states mean · How records are judged · Has this record changed?
Cohere's 111B dense flagship for enterprise RAG and agent workloads with 256K context.
Details
| Released | March 13, 2025stated in cohere.comCommand A was released on March 13, 2025.
|
|---|---|
| Company | Cohere |
| Family | Command |
| Era | Reasoning models |
| Type | Language model · LLM |
| Modalities | Text → Text |
| Capabilities | Agentic, Multilingual, Tool use |
| Weights | Open weights |
| Licence | CC-BY-NC-4.0 |
| Context window | 256K tokensstated in huggingface.coCommand A supports a context window of 256,000 tokens.
|
| Parameters | 111Bstated in huggingface.coCommand A has 111,000,000,000 parameters.
|
Pricing
The price a lab's own page showed on the day that page was captured. That is what a snapshot can prove; it is not necessarily the launch price, and prices are often cut later. Citing the live pricing page instead would give a number with no date at all.
| Input | $2.5 per million tokens observed March 15, 2025 · archived snapshot (2025-03-15) |
|---|---|
| Output | $10 per million tokens observed March 15, 2025 · archived snapshot (2025-03-15) |
Current information, elsewhere
This record is a historical one: every figure above is traced to a dated primary source. For what is true today — live pricing, current rankings, availability — these are better maintained than anything here could be.
- Current price and providers OpenRouter ↗
- Current benchmark scores Artificial Analysis ↗
- Weights and model card Hugging Face ↗
Sources
Record status: verified · confidence 90/100 · 3 of 3 primary
Every value this record asserts — date (Mar 13, 2025), context window (256K) and parameter count (111B) — was found in an archived primary source, checked 2026-08-10. Modalities recorded as text-to-text. The primary source states the model is text-only. Capabilities evidenced from the primary sources: agentic: “com/blog"> < Back to blog Introducing Command A: Max performance, minimal compute Cohere Team Mar 13, 2025 Command A is on par or better than GPT-4o and DeepSeek-V3 ac” | multilingual: “Compared to other leading proprietary and open-weights models Command A delivers maximum performance with minimum hardware costs, excelling on business-critical agentic a” | tool_use: “Tool Use Capabilities: Command A has been specifically trained with conversational tool use capabilities.”.
- cohere.com Official announcement archived
- huggingface.co Model card archived
- cohere.com Official documentation archived
Primary means published by the organisation that made the model. Verified records require at least one.
Command family
- Command R March 11, 2024
- Command A March 13, 2025
Compare Command A with another model →
See Command A on the interactive timeline →