Command A
Command ARankingin non-thinking mode?
Alternatives?
CheaperClaude Haiku 4.5unstated+48%-54%$1 / $5Context: 256K → 200KBetterClaude Opus 4.5unstated+67%+129%$5 / $25Context: 256K → 200K
OpenQwen3 235B A22B Instruct 2507unstated+12%-95%$0.09 / $0.55Memory (4-bit): 67 GB → 141 GB
Cheapero3 (2025-04-16)unstated+36%-20%$2 / $8Context: 256K → 200K
CheaperKimi K2 0711unstatedOpen+27%-80%$0.50 / $2Context: 256K → 128K+2Claude Haiku 4.5unstated+48%-54%$1 / $5Context: 256K → 200Ko3 (2025-04-16)unstated+36%-20%$2 / $8Context: 256K → 200KKimi K2 0711unstatedOpen+27%-80%$0.50 / $2Context: 256K → 128K+2
Gemini 2.5 Flashunstated+21%-81%$0.30 / $2.50GPT-5 Mini (2025-08-07)unstated+19%-84%$0.25 / $2Claude Opus 4.5unstated+67%+129%$5 / $25Context: 256K → 200KClaude Sonnet 4.5unstated+58%+37%$3 / $15Claude Haiku 4.5unstated+48%-54%$1 / $5Context: 256K → 200Ko3 (2025-04-16)unstated+36%-20%$2 / $8Context: 256K → 200KKimi K2 0711unstatedOpen+27%-80%$0.50 / $2Context: 256K → 128K+2Qwen3 235B A22B Instruct 2507unstatedOpen+12%-95%$0.09 / $0.55Memory (4-bit): 67 GB → 141 GBNNanbeige4 3B Thinking 2511Open+11%——Context: 256K → 66KQwen3 32BunstatedOpen+5%-72%$0.70 / $2.80Context: 256K → 33KQwen3 8BunstatedOpen-8%-93%$0.18 / $0.70Context: 256K → 33KQwen3 30B A3B Instruct 2507unstatedOpen-11%-97%$0.09 / $0.30Command R7b-31%—$0.037 / $0.15Context: not verifiedNorth Mini CodeOpenn/a——Command A+thinkingOpenn/a—$0.30 / $1.50Context: 256K → 128K+1Tiny Aya GlobalOpenn/a——Context: 256K → 8K+1Command R7B (12-2024)Openn/a—$0.037 / $0.15Context: 256K → 128K+1Command R (08-2024)unstatedOpenn/a—$0.15 / $0.60Context: 256K → 128K+1Command R+ (08-2024)unstatedOpenn/a—$2.50 / $10Context: 256K → 128K+1Command R+Openn/a——Context: 256K → 128KAll Cohere models →Run it yourself
estimateBenchmarks
all categoriesFAQ
How much does Command A cost?›
$2.50 per 1M input tokens and $10 per 1M output tokens.
Is Command A free to use?›
There is no free API; the cheapest paid price is $2.50 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Command A, and what is it best at?›
#307 of 334 in Chat & Knowledge, #203 of 239 in Coding, #209 of 242 in Agents & Tool Use, #110 of 195 in Writing & Language.
Is Command A better than Qwen3 32B?›
On BFCL v4, Command A scores 46.5 vs 48.7 — lower.
How can I use Command A?›
Through the API of Azure AI Foundry, Cohere.
What is the context window of Command A?›
256K tokens, with up to 8K output tokens.
When was Command A released? Is there a newer version?›
Released 2025-03-13. It is the current version in its line.
Is Command A open source?›
Yes, the weights are open (cc-by-nc-4.0 licence).
Can I run Command A locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Command A read images?›
No, it accepts text only.
What is Command A?›
Command A is an open-weights language model from Cohere, released Mar 2025.
What are the best alternatives to Command A?›
Claude Haiku 4.5 (cheaper), Claude Opus 4.5 (better), Qwen3 235B A22B Instruct 2507 (open weights).