Command A

Frontier

Cohere•Released on 2025-03-15

Cohere's most performant Command model. 150% throughput of Command R+ on only 2 GPUs. Enterprise-optimized for RAG and agentic tasks. 256K context. Strong for business workflows.

80
Overall Score

Core Specs

256K
Context Window
16K
Max Output
✗
Reasoning
✗
Open Source
Multimodal Support
text

Scenario Scores

Coding
#21 / 2478
Writing
#18 / 2375
Translation
#16 / 2380
Data Analysis
#11 / 2485
Conversation
#17 / 2380
Image Gen
—

User Feedback Highlights

Based on community feedback. Hover to see original reviews.

+ Efficient GPU usage− Higher price than competitors+ Enterprise-optimized for RAG+ 150% throughput improvement+ Strong agentic capabilities+ 256K context window− No multimodal support− Less known in developer community− Smaller ecosystem than OpenAI/Anthropic
Sentiment:👍 68%😐 25%👎 7%

Pros & Cons

Pros

  • +Enterprise-optimized for RAG
  • +150% throughput improvement
  • +256K context window
  • +Strong agentic capabilities
  • +Efficient GPU usage

Cons

  • −Higher price than competitors
  • −Smaller ecosystem than OpenAI/Anthropic
  • −No multimodal support
  • −Less known in developer community

Reliability

SLA99.9%
Incidents (30d)0
View Status Page →

Pricing

Input (per 1M tokens)$2.50
Output (per 1M tokens)$10.00
Free trial available
Updated on 2026-03-07

Compare with Others

Command A vs GPT-5.4 →Command A vs GPT-5.4 Thinking →Command A vs Claude Opus 4.6 →

Benchmarks

mmlu86.5%
humanEval85%