Gemini 4 Argon
Claude Opus 5.5
Muse Wonder
MiMo-V2.6-Pro
Qwen3.8-Max
Kimi K3
GLM-5.3
GPT-5.5
ERNIE 5.1
DeepSeek-V4.1-Flash-20260910
Grok 4.20 Beta
amazon-nova-experimental-chat
Dola Seed 2.0 Pro
NVIDIA-Nemotron-3-Ultra-550B-A55B
Gemma 4 31B
Step 5 Preview
Command A
Command A is an open-weight model from Cohere, released Sep 22, 2026. It ranks #227 of 351 on the LMArena text leaderboard with a rating of 1331. It costs $0.30 per million input tokens and $1.50 per million output tokens, with a 192K-token context window.
- Developer
- Cohere
- Released
- Sep 22, 2026
- Context window
- 192K
- Price in / out, per 1M tokens
- $0.30 / $1.50
- Weights
- Open · CC-BY-NC-4.0
- Inputs
- Text, Images
- Reasoning
- Yes
- Tool use
- Yes
Benchmarks
| Benchmark | Score bar | Accuracy | Rank | Cost / test |
|---|---|---|---|---|
| SWE-benchCoding | 7.80% | 85 / 85 | $0.21 | |
| LiveCodeBenchCoding | 35.07% | 137 / 143 | $2.5 / $10 | |
| MedCodeMedicine | 19.72% | 107 / 107 | N/A | |
| MedQAMedicine | 80.55% | 70 / 92 | $2.5 / $10 | |
| GPQA DiamondReasoning | 48.48% | 125 / 138 | $2.5 / $10 | |
| MMLU-ProReasoning | 69.17% | 125 / 138 | $2.5 / $10 | |
| AIMEReasoning | 13.33% | 86 / 96 | $2.5 / $10 |
Vals AI, as of Oct 24, 2026.
LMArena, by category
| Leaderboard | Rating | Rank | Votes |
|---|---|---|---|
| TextLMArena | 1331 | 227 / 351 | 55.8K |
| CodingLMArena | 1330 | 237 / 346 | 10.2K |
| MathLMArena | 1300 | 242 / 339 | 4K |
| Creative writingLMArena | 1319 | 201 / 349 | 7.6K |
| Instruction followingLMArena | 1309 | 228 / 351 | 15.6K |
| Hard promptsLMArena | 1326 | 231 / 351 | 23.8K |
| Multi-turnLMArena | 1338 | 217 / 349 | 9.3K |
| Long queriesLMArena | 1334 | 211 / 330 | 10.7K |
| ExpertLMArena | 1296 | 235 / 305 | 2.9K |
| Legal and governmentIndustries | 1348 | 211 / 325 | 3.4K |
| Medicine and healthcareIndustries | 1334 | 225 / 325 | 3.1K |
| Business and financeIndustries | 1315 | 228 / 344 | 8.8K |
| ChineseLanguages | 1327 | 234 / 328 | 3.8K |
| SpanishLanguages | 1347 | 198 / 249 | 1.2K |
| GermanLanguages | 1341 | 177 / 261 | 1.4K |
| FrenchLanguages | 1351 | 193 / 252 | 776 |
| JapaneseLanguages | 1285 | 172 / 236 | 1.2K |
LMArena, CC BY 4.0, as of Oct 22, 2026. Where a model is rated at several effort levels, its best is shown.
Compare Command A
Cohere
An enterprise-focused lab behind Command models, retrieval and embeddings. Founded 2019, based in Toronto. cohere.com
Questions
Who makes Command A?
Command A is made by Cohere, founded 2019 and based in Toronto.
How much does Command A cost?
$0.30 per million input tokens and $1.50 per million output tokens, at its provider’s list price.
What is Command A’s context window?
192K tokens (192,000).
How does Command A rank?
#227 of 351 on LMArena’s text leaderboard, rated 1331 (95% interval 1328–1334) from 55,758 votes, as of Oct 22, 2026.
Is Command A open source?
Command A has open weights under the CC-BY-NC-4.0 license.