COHERE · MODEL RELEASE TRACKER

Command A+

Command A+ is referenced as one of Cohere's own models. The only supplied source reports it as supported by Cohere's North 2 enterprise platform, which is described as model-agnostic and able to orchestrate agents for multi-step workflows, tool integrations, and session-level context retention.

UPDATED 2026-10-06

CURRENT SNAPSHOT3/5 DIMENSIONS WITH DATA

The dimensions that change the decision.

PRICING

Not established from the available sources.

CONTEXT WINDOW

Not established from the available sources.

MODALITIES

ModalityMultimodal and multilingual large language model (LLM).2026-09-16

BENCHMARKS

BenchmarksReported to outperform previous Command A models in multimodal understanding, retrieval, long-horizon, and complex reasoning; τ²-Bench Telecom score improved from 37% (Command A Reasoning) to 85%.2026-09-16

AVAILABILITY

Platform supportReported as one of Cohere's own models available to plug into North 2's model-agnostic enterprise platform.2026-10-06

RELEASE TIMELINE

Published, source-backed release events only.

2026-10-06

1 SOURCES · IMPORTANCE 7.0

Cohere launches North 2 — model-agnostic enterprise AI control room

Cohere released North 2, an enterprise platform that acts as a control center for AI agents and multi-step workflows, retaining session context and coordinating agents via a new orchestration system that taps shared knowledge libraries and reusable skills. North 2 is model-agnostic (including Cohere’s Command A+ and customer models), can run on-premises, in cloud, or air-gapped, integrates with tools like Slack, SharePoint, and Jira, offers admin controls via “North Admin,” requests human approval for critical agent actions, and — Cohere says — runs more token-efficiently on Nvidia Blackwell and Hopper; LG CNS and Bell Cyber are cited as production users.

→

VERIFIABLE FACTS

Every value stays attached to a source and date.

RELEASE · ReleaseDEVELOPER CLAIM

Released open-source on 2026-09-16; weights and implementation guides published and distributed under an Apache 2.0 license.

CAPABILITIES · CapabilitiesDEVELOPER CLAIM

Described as Cohere’s fastest and most powerful language model to date; optimized for complex reasoning, multimodal and multilingual agentic tasks; optimized for low-bit quantization, efficient inference, and integration with open inference frameworks.

MODALITIES · ModalityDEVELOPER CLAIM

Multimodal and multilingual large language model (LLM).

CAPABILITIES · ArchitectureDEVELOPER CLAIM

Mixture-of-experts (MoE) model.

CAPABILITIES · Minimum computeDEVELOPER CLAIM

Can run on as little as two NVIDIA H100 GPUs.

AVAILABILITY · Platform supportDEVELOPER CLAIM

Reported as one of Cohere's own models available to plug into North 2's model-agnostic enterprise platform.

BENCHMARKS · BenchmarksDEVELOPER CLAIM

Reported to outperform previous Command A models in multimodal understanding, retrieval, long-horizon, and complex reasoning; τ²-Bench Telecom score improved from 37% (Command A Reasoning) to 85%.

CAPABILITIES · Agent orchestration compatibilityDEVELOPER CLAIM

Reported to be usable within North 2's orchestration of agents that handle multi-step workflows, connect to tools (e.g., Slack, SharePoint, Jira), and retain context across sessions.

WHAT CHANGED

Stored passport versions, without reconstructed history.

2026-10-06Passport updated8 facts

AVAILABILITY · Availability & deployment− Weights available on Hugging Face in several near-lossless quantizations; can be deployed in Cohere’s managed Model Vault for inference; released for private deployment by developers.2026-09-16

CAPABILITIES · Agent orchestration compatibility+ Reported to be usable within North 2's orchestration of agents that handle multi-step workflows, connect to tools (e.g., Slack, SharePoint, Jira), and retain context across sessions.2026-10-06

AVAILABILITY · Platform support+ Reported as one of Cohere's own models available to plug into North 2's model-agnostic enterprise platform.2026-10-06

2026-09-29Passport created7 facts

Cohere launches North 2 — model-agnostic enterprise AI control room

Released open-source on 2026-09-16; weights and implementation guides published and distributed under an Apache 2.0 license.

Described as Cohere’s fastest and most powerful language model to date; optimized for complex reasoning, multimodal and multilingual agentic tasks; optimized for low-bit quantization, efficient inference, and integration with open inference frameworks.

Multimodal and multilingual large language model (LLM).

Mixture-of-experts (MoE) model.

Can run on as little as two NVIDIA H100 GPUs.

Reported as one of Cohere's own models available to plug into North 2's model-agnostic enterprise platform.

Reported to outperform previous Command A models in multimodal understanding, retrieval, long-horizon, and complex reasoning; τ²-Bench Telecom score improved from 37% (Command A Reasoning) to 85%.

Reported to be usable within North 2's orchestration of agents that handle multi-step workflows, connect to tools (e.g., Slack, SharePoint, Jira), and retain context across sessions.

Optional analytics