Command A+
Cohere · released
Cohere's flagship: an Apache-2.0 open-weights model that folds reasoning, image input, tool use and 48-language support into one model sized for one or two data-centre GPUs. It is aimed at enterprise agents and retrieval in private deployments, not at the top of the leaderboards.
Best for: Private enterprise deployment · Retrieval and tool use · Multilingual work
- Intelligence Index
- 13.1
- #36 of 38
- Price per 1M tokens
- $0.30 in $1.5 out
- Blended $0.60 at 3:1
- Context window
- 128K
- Up to 64K output
- Output speed
- 159 tokens/s
- Median, measured by Artificial Analysis
Capability profile
One 0–100 score per area, placing this model between the weakest and strongest tracked model across that area's benchmarks.
- Overall4 · #36 of 38
Based on 1 of 3 benchmarks
- Coding5 · #36 of 39
Based on 2 of 5 benchmarks
- Agents & knowledge work1 · #38 of 40
Based on 1 of 6 benchmarks
- Reasoning & science3 · #37 of 39
Based on 2 of 2 benchmarks
- Knowledge & reliability36 · #34 of 39
Based on 2 of 2 benchmarks
- Vision21 · #29 of 31
Based on 1 of 2 benchmarks
Benchmark results
All third-party results on record. Rank is among the tracked models that have been tested on that benchmark.
Overall
#36 of 38· Artificial Analysis
- Vals IndexNot tested
- Arena · TextNot tested
Coding
#31 of 38· Artificial Analysis
- SciCode38.5%
#37 of 38· Artificial Analysis
- Arena · WebDevNot tested
- Vibe Code BenchNot tested
- Code MigrationNot tested
Agents & knowledge work
- GDPval-AA0.9
#37 of 39· Artificial Analysis
- Arena · AgentNot tested
- Finance Agent v2Not tested
- Excel Modeling BenchmarkNot tested
- Legal Research BenchNot tested
- Tax Agent BenchNot tested
Reasoning & science
- Humanity's Last Exam12.0%
#37 of 39· Artificial Analysis
- CritPt0.3%
#35 of 39· Artificial Analysis
- ARC-AGI-2saturatingNot tested
- ARC-AGI-3Not tested
- GPQA Diamondsaturating76.1%
#26 of 27· Artificial Analysis
Knowledge & reliability
#25 of 39· Artificial Analysis
- AA-LCR52.7%
#37 of 39· Artificial Analysis
Vision
- MMMU-Pro63.2%
#26 of 27· Artificial Analysis
- Arena · VisionNot tested
Reported by Cohere
Numbers the lab published at launch, run on its own harness and settings. Useful context, but not independently verified and never used in the rankings on this site.
- τ²-Bench Telecom85%
- MMMU75.1%
- Terminal-Bench Hard · Terminus-2 harness25%
Specifications
- Lab
- Cohere
- Released
- May 20, 2026
- Availability
- Generally available
- Context window
- 128,000 tokens
- Max output
- 64,000 tokens
- Accepts
- Text, Images
- Knowledge cutoff
- Not published
- Weights
- Open
- Licence
- Apache 2.0
- Parameters
- 218B total · 25B active
- Reasoning control
- Thinking on or off, with a token budget
- API model ID
command-a-plus-05-2026
Pricing
- Input / 1M tokens
- $0.30
- Cached input / 1M
- $0.15
- Output / 1M tokens
- $1.5
- Blended (3:1)
- $0.60
Cohere lists rate-limited API access as free and prices production use through sales. This is the per-token rate of Cohere's own endpoint on OpenRouter.
Official sources
Closest alternatives
The models nearest to this one on overall score.
23 benchmarks tracked. See methodology and sources.