ModelBeat
Models/Command A+
All models
C

Cohere,Cohere Labs (formerly Cohere for AI): Command A+

cohere/command-a

Command A+ is Cohere,Cohere Labs (formerly Cohere for AI)’s open-weights AI model, released May 20, 2026. It costs $2.50/M input and $10/M output tokens, and handles a 256K-token context window.

Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding use cases.

Context
256K
Input
$2.50 / 1M
Output
$10 / 1M
Released
May 20, 2026
Modalities
texttext

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Command A+ lands among all models tracked on Model Beat.

21.0
Intelligence Index
Artificial Analysis
21st percentile of tracked models
20.0
Coding Index
Artificial Analysis
20th percentile of tracked models
29.0
Agentic Index
Artificial Analysis
29th percentile of tracked models

Reasoning

2 evals
GPQA Diamond76.1%

Graduate-level scientific reasoning

Humanity's Last Exam12.0%

Frontier of human expert knowledge

Coding

1 evals
SciCode38.5%

Scientific research coding

Agentic & Tools

1 evals
τ²-bench80.7%

Tool-agent-user reliability

In the news

6

Changelog

1
  • Command A+: Humanity's Last Exam improved from 11.4% to 12.0% (+5%)benchmark · Aug 6, 2026

Compare Command A+ with

Frequently asked questions

What is Command A+?

Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding use cases.

Who created Command A+?

Command A+ is an AI model developed by Cohere,Cohere Labs (formerly Cohere for AI), part of the Command family, released May 20, 2026.

How much does Command A+ cost?

Command A+'s list price from Cohere,Cohere Labs (formerly Cohere for AI) is $2.50 per million input tokens and $10 per million output tokens.

What is Command A+'s context window?

Command A+ supports a context window of up to 256K tokens.

How does Command A+ perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Command A+ scores 80.7% on τ²-bench, 38.5% on SciCode, 76.1% on GPQA Diamond.

Is Command A+ open source?

Command A+ is released with open weights (Open weights (unrestricted)), so the model can be downloaded and self-hosted.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Sep 10, 2026.