Command A+
Cohere's open-weight enterprise flagship — a sparse MoE built for RAG, tool use, and sovereign deployment.
About
Command A+ (model ID command-a-plus-05-2026) is Cohere's flagship: a decoder-only sparse MoE with 218B total / 25B active parameters, reasoning, vision input, 48-language support, and structured outputs with citations.
It is designed for enterprise agentic workflows and sovereign AI — private and air-gapped deployment are supported, and weights are Apache 2.0 licensed. It is the default model in Cohere's API reference examples.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Benchmark comparisons vs Command A Reasoning on multimodal and agentic coding tasks reported by Cohere (provider-claimed, per independent coverage). No independently verified scores found.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Cohere is an established enterprise provider, specs are clearly documented, and Apache 2.0 weights with air-gapped deployment are a real differentiator for regulated industries. But pricing is undisclosed (sales-gated), benchmark evidence is thin and entirely vendor-reported, and the 128K context is the smallest among the catalog entries — fine for RAG, limiting for repo-scale agents.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.7/10
- Maintenance · recency, releases, issue load
- 9.0/10
- Openness · license + docs
- 10.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · HF top match: CohereLabs/command-a-plus-05-2026-bf16 (community upload counts as proxy where official repo is gated)
- · HF downloads 22,144 (peer-normalized)
- · released 11 days ago
- · manual review: undisclosed pricing; thin public benchmarks