Command A — API Pricing & Benchmarks
Command A is Cohere's mid model, released in 2025-03. The API costs $2.50 per 1M input tokens and $10.0 per 1M output tokens ($4.38 blended at a 3:1 ratio). It supports a 256K-token context window. On the independently measured Artificial Analysis Intelligence Index it scores 7.5, ranking #66 of 68 models tracked on this site, with a median output speed of 57 tokens/sec.
Explore provider routes and additional benchmark sources for Command A
| Provider | Cohere |
| Input price / 1M tokens | $2.50 |
| Output price / 1M tokens | $10.0 |
| Cached input / 1M tokens | — |
| Blended price (3:1) | $4.38 |
| Context window | 256K |
| Max output tokens | 8K |
| Knowledge cutoff | 2024-06 |
| License | open weights |
| Input modalities | T |
| Output speed (tok/s) | 57 |
| Released | 2025-03 |
| Artificial Analysis Intelligence Index | 7.5 |
| GPQA Diamond | 52.7% |
| Humanity's Last Exam | 4% |
Where Command A fits
On a blended 3:1 basis Command A costs $4.38 per 1M tokens, cheaper than 24% of the 80 priced models in this index. It sits in the mid-tier bracket where most production chat, RAG and summarization workloads run: meaningfully cheaper than flagships while keeping most of their capability. Measured output speed of 57 tokens/sec is below the index median of 104 — typical of reasoning-heavy models, and worth checking against interactive latency budgets. The weights are open, so it can be self-hosted for data-residency or cost reasons instead of consumed via API.
Price history
| Period | Input $/1M | Output $/1M | Cached $/1M |
|---|---|---|---|
| 2026-07-18 → current | $2.50 | $10.0 | — |
No list-price changes recorded since tracking began. Full dataset: price-history.json (CC BY 4.0).
Alternatives at a similar price
- GPT-4o (OpenAI) — $4.38/1M blended, AA Index 12.3
- GPT-5.6 Terra (OpenAI) — $4.50/1M blended, AA Index 56.6
- Gemini 3.1 Pro (Google) — $4.50/1M blended, AA Index 47.7
- Claude Sonnet 5 (Anthropic) — $4.00/1M blended, AA Index 55.3
About Cohere
Cohere targets enterprise deployments with its Command series, emphasizing retrieval-augmented generation, tool use and private-deployment options over leaderboard placement.