Command A
Released 2025-03 · chat · 256K tokens · 5 benchmarks · Open weight
Editorial notes
111B params, corre en 2 GPUs. 156 tokens/sec (1.75x GPT-4o). Enfoque enterprise: tools, RAG, agents. Superseded por Command A+ (20-may-2026), que unifica Command A + Command A Reasoning. Los 5 scores confirmados contra el paper oficial arxiv 2504.00698 (Tables 6 y 15); GPQA es Diamond subset, MATH es MATH-all.
Spec sheet
- Company
- Cohere
- Country
- CA
- Type
- chat
- Release
- 2025-03
- Context
- 256K tokens
- Params total
- 111B
- License
- cohere-non-commercial
- Quants
- BF16, Q8_0, Q4_K_M
- Pricing (cohere)
- $2.5/$10/M
- Slug
- command-a
Benchmarks (5)
Reasoning 2
Cite this model
BibTeX · APA
BibTeX
@misc{frontier-command-a,
title = {Command A},
author = {{Cohere}},
year = {2025},
note = {Frontier Benchmarks AI atlas. Accessed 2026-06-17},
url = {https://frontierbenchmarks.com/models/command-a}
} APA
Cohere (2025). Command A [Large language model]. Frontier Benchmarks AI. Retrieved 2026-06-17, from https://frontierbenchmarks.com/models/command-a
Citation reflects the atlas page, not the original model paper. For the paper, see the "Resources" section above.