Cohere model product
Command A+
Command A+ is Cohere's open-source sparse Mixture-of-Experts model for enterprise agentic, multilingual, multimodal, and reasoning-heavy tasks.
Updated Oct 23, 2026. Default version: Command A+
LLMBoard Score39.6Command A+
Coverage60%10 benchmark families
Context windowN/ATokens
Official input price$2.5Cohere API
Command A+ Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Command A+ Benchmark Results
Benchmark scores for Command A+.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkCohere Agentic Question Answering | Score65.00% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkCohere Data Analysis | Score45.00% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkCohere Memory Usage Quality | Score54.00% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkMT-AIME 2025 | Score86.00% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkTerminal-Bench Hard | Score25.00% | Rank02 | Participants2 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkMathVista | Score80.60% | Rank07 | Participants40 | Percentile84.62% | EvidenceC | Evaluated |
| BenchmarkWMT24++ | Score81.00% | Rank07 | Participants24 | Percentile73.91% | EvidenceC | Evaluated |
| BenchmarkCharXiv-D | Score88.00% | Rank08 | Participants18 | Percentile58.82% | EvidenceC | Evaluated |
| BenchmarkTau2 Telecom | Score85.00% | Rank20 | Participants37 | Percentile47.22% | EvidenceC | Evaluated |
| BenchmarkSciCode | Score38.00% | Rank23 | Participants31 | Percentile26.67% | EvidenceC | Evaluated |
| BenchmarkMMMU | Score75.10% | Rank24 | Participants66 | Percentile64.62% | EvidenceC | Evaluated |
| BenchmarkIFBench | Score74.00% | Rank25 | Participants46 | Percentile46.67% | EvidenceC | Evaluated |
| BenchmarkAA IFBench | Score73.95% | Rank26 | Participants168 | Percentile85.03% | EvidenceB | EvaluatedN/A |
| BenchmarkAIME 2025 | Score90.00% | Rank50 | Participants122 | Percentile59.50% | EvidenceC | Evaluated |
| BenchmarkMMMU-Pro | Score63.00% | Rank50 | Participants73 | Percentile31.94% | EvidenceC | Evaluated |
| BenchmarkCharXiv-R | Score52.70% | Rank55 | Participants60 | Percentile8.47% | EvidenceC | Evaluated |
| BenchmarkAA SciCode Subtasks | Score38.54% | Rank67 | Participants89 | Percentile25.00% | EvidenceB | EvaluatedN/A |
| BenchmarkAA CritPt | Score0.29% | Rank113 | Participants200 | Percentile43.72% | EvidenceB | EvaluatedN/A |
| BenchmarkAA HLE Text No Tools | Score11.96% | Rank117 | Participants200 | Percentile41.71% | EvidenceB | EvaluatedN/A |
| BenchmarkAA GPQA Diamond | Score76.06% | Rank119 | Participants199 | Percentile40.40% | EvidenceB | EvaluatedN/A |
| BenchmarkAA LCR v1.1 | Score52.67% | Rank124 | Participants193 | Percentile35.94% | EvidenceB | EvaluatedN/A |
| BenchmarkAA Omniscience Accuracy | Score8.88% | Rank192 | Participants201 | Percentile4.50% | EvidenceB | EvaluatedN/A |
Command A+ Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
Command A+ Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $2.5 input, $10 output per 1M
- Official provider
- Cohere
- Lowest third-party
- N/A
- Tracked offerings
- 1
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderCohere | Provider model IDcommand-a-plus-05-2026 | Regionglobal | Input / 1M$2.5 | Output / 1M$10 | Context128K | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Command A+ Runtime Performance
Provider-specific output speed and catalog latency for Command A+. Runtime does not affect the capability score.
No runtime data
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Command A+ Specifications
Technical details for the model's default version.
- Version
- Command A+
- Released
- May 20, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 218B
- Context window
- N/A
- Max output
- N/A
- Inputs
- image, text
- Outputs
- text
- Open weights
- Yes
- License
- Apache 2.0
Command A+ Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionCommand A+ | Released | LLMBoard39.59 | Parameters218B | ContextN/A | Max outputN/A | Open weightsYes | LicenseApache 2.0 |
Models similar to Command A+
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
CO
North Mini Code 1.0
Cohere
CO
North Micro Vision
Cohere
CO
Command R+
Cohere
AC
Qwen3.5 9B
Alibaba Cloud / Qwen Team
ZA
GLM 4.5 Air
Zhipu AI
PO
Laguna XS 2.1
Poolside
What is Command A+?
Key information about Command A+ and its available data.
Command A+ is a Cohere large language model with 218B total parameters and 25B active parameters. It supports text and image inputs, conversational tool use, generated thinking, 48 languages, a 128K input context, and up to 64K output tokens.
Data as of 2026-10-22.
FAQ
Common questions about Command A+.
When was Command A+ released?
Command A+'s default version was released on May 20, 2026.
How much does Command A+ cost?
Command A+'s official API price is $2.5 per million input tokens and $10 per million output tokens via Cohere.
Who created Command A+?
Command A+ was created by Cohere.
What is the context window for Command A+?
A context window is not available for the default version.
Is Command A+ open weight?
Yes. The default version is marked as open weight under Apache 2.0.
How many API providers offer Command A+?
1 provider offerings are linked to the default version.