Command A Reasoning (08-2025)
Released
2025-08-01
Last refreshed
2026-10-02
Status
Researched 43d ago
Open weightsCommercial use: not permittedRAGAgentsLong contextJSON / Tool use
Command A Reasoning (08-2025) is a released rag, agents, and long context model with open-weight and 256K context; evaluate it while provider pricing coverage matures.
Cohere releases · 10 in the last 12 monthsChangelog →
Specifications
- Family
- Command
- Released
- 2025-08-01
- Context
- 256K
- Architecture
- Decoder Only
- Specialization
- reasoning
- Openness
- Open weights
- License
- CC-BY-NC-4.0Not permitted
- Weights
- Unknown
- Code
- Unknown
Created by
Pricing
No tracked provider pricing is available yet.
About
Cohere's first reasoning model capable of 'thinking' before generating an output. Command A Reasoning allows nuanced problem-solving and agent-based tasks in 23 languages. Supports up to 32K output tokens for detailed reasoning and analysis.
Providers
No active provider routes are tracked for this model yet.
Capabilities
JSON / Tool use
Benchmark peer barsfor JSON / Tool use
BFCLRank 12 of 58
Benchmark scores(1)
Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
| Benchmark | Score | Version | Evaluation | Source |
|---|---|---|---|---|
| 57.1 | v4Observed 2026-10-26 | Configuration: Command A Reasoning (FC) Harness: BFCL V4 official evaluation (gorilla berkeley-function-call-leaderboard), FC mode Evaluator: Gorilla LLM team (UC Berkeley) Confidence: confirmed Notes: BFCL V4 leaderboard (Last Updated 2026-04-12), entry 'Command A Reasoning (FC)': rank 13, Overall Acc 57.06%; Multi Turn 50.12%, Web Search 55.50%, Memory 28.82%, Non-Live AST 86.27%, Live 78.61%, Irrelevance 86.75%. Retrieved 2026-10-26. |
Migration checks
No linked migration route is available for this model yet.