Google gives Gemini a face: Live Avatar with real-time lip sync and video conversations in 97 languages

From Voice Assistant to Voice AgentCohere Hearing Enhancement Command A+Meta calls this “FDA-cleared software for mild to moderate hearing loss,” launching this year as part of the Meta One subscription or as a standalone add-on for $149.

core competencies

A new heavyweight has entered the enterprise AI model race. On September 22,

• Cohere launched its next-generation flagship model,designed specifically for enterprise Agent workflows, supporting text and image multimodal input, 192K ultra-long context, native tool calling, and structured output; API pricing is $0.30 per million tokens for input and $1.50 for output. For teams evaluating AI models to get real work done in enterprise settings, this is a compelling new option worth serious consideration.
• Command A+ is Cohere’s flagship model built for enterprise Agentic Workflow—its positioning is clear: it is not a general-purpose chat companion but a productivity tool built to get work done. Its core capabilities are centered on enterprise automation:Text + image multimodal input
• It processes not only text but also understands images—ideal for document processing, invoice recognition, and report analysis.192K context window
• Supports ingestion of extremely long documents in a single pass—capable of handling lengthy reports, multi-turn conversations, and codebase-scale tasks.Native tool calling
• Enforces strict tool schema constraints, enabling reliable invocation of external APIs, databases, and internal systems.Structured output
• Supports JSON Object and JSON Schema, delivering predictable, format-controlled outputs that integrate seamlessly with downstream systems.Optional reasoning mode

From “Voice Assistant” to “Voice Agent” Upgrade

Hearing Enhancement Meta calls this “FDA-cleared software for mild to moderate hearing loss,” launching this year as part of the Meta One subscription or as a standalone add-on for $149.Pricing and Positioning

Command A+ pricing is

$0.30 per million input tokens and $1.50 per million output tokensIn industry context, this is significantly cheaper than OpenAI’s and Anthropic’s top-tier flagship models, yet more expensive than open-source self-hosted models—it squarely targets the enterprise mid-market segment seeking flagship capabilities without flagship pricing. Combined with Cohere’s consistent “private deployment / VPC isolation” capability, enterprises with strict data compliance requirements can run it in their own environments.What use cases is it suited for?Command A+’s typical enterprise automation use cases include:Intelligent customer support and ticketing systems(multimodal understanding of screenshots and attachments + strictly structured output integrated with ticketing systems),Document and contract processing(192K context window enabling full-contract and invoice ingestion in one pass, outputting structured fields),Internal knowledge base Q&A agents

(tool calling connected to databases and internal systems, enabling on-demand reasoning for complex queries), and

User experience/limitations

Private deployments for compliance-sensitive industries(finance, healthcare, and government sectors with hard requirements for data residency).From Voice Assistant to Voice Agent

Hearing EnhancementMeta calls this “FDA-cleared software for mild to moderate hearing loss,” launching this year as part of the Meta One subscription or as a standalone add-on for $149.These three capabilities represent the most essential requirements for enterprise Agent deployment—and Cohere has consistently demonstrated strong engineering expertise in precisely these areas. Its 192K context window and configurable safety mode further widen the gap between it and “toy-grade” models. ChatGPT Limitations are also evident:

It is not a consumer-facing product.

Model	context	Core Positioning	This Site’s Rating
Cohere Command A+	192K	There is no C-end entry point—making it cumbersome for individual users and content creators to write articles, generate images, or chat. Its ecosystem and community momentum also pale in comparison to OpenAI or Google. Moreover, as a new model, its reasoning ceiling and performance in Chinese-language scenarios still require validation through more third-party benchmarks and real-world business data.	8.4
Claude Opus 5.5	200K+	Command A+ vs. Other Comparable Models	9.1
GPT-6 Sol	200K+	Enterprise Agent Workflow	9.0
Qwen3.8 2.4T	1M	Safety-Aligned Flagship	8.8
GLM 5.3 FlashX	200K	Efficiency-Oriented Flagship	8.6

Open-Weight Giant Claude Domestic High-Speed Multimodal Model

Overall Score

维度	Score	evaluate
functional completeness	8.5 / 10	Geographically speaking, Command A+’s overall capability slightly trails top-tier flagships like Opus 5.5 and GPT-6—but its differentiation lies in “enterprise compliance + tool calling + configurable safety,” and its pricing is more accessible. If you seek the best-value enterprise Agent foundation, Command A+ delivers better ROI than blindly opting for top-tier specs.
易用性	8.0 / 10	Fully equipped with multimodality, tool calling, structured output, and a 192K context window
Cost-effectiveness	8.5 / 10	From Voice Assistant to Voice Agent Upgrade
中文支持	7.5 / 10	Hearing Enhancement
输出质量	8.5 / 10	Meta calls this FDA-cleared software for mild to moderate hearing loss; launching this year, available via Meta One subscription or as a $149 standalone add-on.

Overall rating: 8.4/10

Frequently Asked Questions (FAQ)

Priced below top-tier flagships, strong competitiveness in the mid-tier segment

Supports multiple languages, but Chinese is not a strength—further validation neededMeta calls this “FDA-cleared software for mild to moderate hearing loss,” launching this year as part of the Meta One subscription or as a standalone add-on for $149.Stable in enterprise scenarios, with reliable structured output

How much does Cohere Command A+ cost?

API pricing is

Already live on platforms including OpenRouter, and also accessible via Cohere’s official API and private deployment. ChatGPT,Claude Does Command A+ support Chinese?

Supports multilingual input and output; Chinese works normally, but Cohere’s strength lies in English enterprise use cases—Chinese performance should be validated against your specific tasks.What’s the difference between Command A+ and?ChatGPT,Claude Different positioning. Command A+ is a model designed for enterprise Agent workflows, emphasizing tool calling, structured output, and configurable security;

is an assistant geared toward individual and general-purpose use, with stronger ecosystem integration and consumer-facing experience.

Can Command A+ be deployed privately?

Want to Discover More Useful AI Tools? Explore Our AI Model Library and Tool Comparison Engineor continue reading:Claude Opus 5.5 Deep Review · Yes. Cohere offers private deployment and VPC isolation solutions, suitable for enterprises with strict data compliance requirements. · From “Voice Assistant” to “Voice Agent”: An Upgrade · DeepSeek V4 Pro In-Depth Evaluation.