Activefrontierllm Proprietary

command-a-plus-05-2026

by Cohere· Released May 2026· Cutoff April 2026

Command A+ is Cohere's most advanced large language model, optimized for enterprise-grade reasoning, long-context understanding, and multilingual tasks. It builds on the Command family with enhanced accuracy, lower latency, and stronger instruction following. This model is designed for complex workflows requiring high reliability and safety.

Official Site API Docs

Input cost

$2.50 per 1M tokens

Output cost

$10.00 per 1M tokens

Context window

256K tokens

Max output

8192 tokens

Modalities

text

License

proprietary

Capabilities

Function CallingCode GenerationStreamingJSON ModeMultilingual SupportLong Context UnderstandingRetrieval Augmented Generation (RAG)

Best For

Enterprise applications requiring high-accuracy reasoning, long document analysis, and multilingual support.

Strengths

State-of-the-art reasoning and factual accuracy
256K token context window for processing large documents
Strong multilingual performance across 100+ languages
Low latency suitable for real-time applications
Robust safety and alignment features

Limitations

No native vision or image input support
Higher cost per token compared to smaller models
May still produce hallucinations on niche topics
Not optimized for creative writing or open-ended generation

Use Cases

Enterprise document analysis and summarization

Multilingual customer support chatbots

Code generation and debugging assistance

Legal and financial contract review

Knowledge base question answering with RAG

Data extraction and structured output generation

Complex reasoning tasks like math and logic

Improvements Over Previous Model

Context window increased from 128K to 256K tokens
Improved accuracy on reasoning benchmarks (e.g., +5% on MMLU)
Lower latency: ~2x faster inference than Command A
Enhanced multilingual support with better performance on low-resource languages
Reduced pricing: $2.50/M input tokens vs $3.00/M for Command A
Improved instruction following and reduced hallucination rates

Back to all models

Improvements Over Previous Model

Context window increased from 128K to 256K tokens

Improved accuracy on reasoning benchmarks (e.g., +5% on MMLU)

Lower latency: ~2x faster inference than Command A

Enhanced multilingual support with better performance on low-resource languages

Reduced pricing: $2.50/M input tokens vs $3.00/M for Command A

Improved instruction following and reduced hallucination rates