North Mini Code (free)

Overview

North Mini Code is Cohere’s first agentic coding model and the first release in its North family. It is a sparse mixture-of-experts model with 30 billion total parameters and 3 billion active parameters. That distinction describes the model’s overall and active parameter counts; it does not, by itself, establish how it will perform on a particular coding workload.

The model accepts text and returns text, with reasoning and tool support. Its 256,000-token context window gives it room to take in substantial text input, while its maximum output is 64,000 tokens. Those limits may matter for workflows that pass large code-related contexts or request extended responses, but the listed specifications do not indicate which programming languages, repositories, or tools it handles.

North Mini Code is free, and its weights are open. The listed model identifier is cohere/north-mini-code:free; its Hugging Face identifier is CohereLabs/North-Mini-Code-1.0. It was released on June 17, 2026.

Key features

Agentic coding focus

Cohere positions North Mini Code as an agentic coding model. It supports tools, a relevant capability for workflows that coordinate model responses with external actions. The available specifications do not name specific integrations or describe an execution environment, so teams should confirm that their intended tools and orchestration layer are supported before adopting it.

Context and output capacity

The 256,000-token context window and 64,000-token maximum output are the model’s stated limits. These provide a substantial text-processing envelope on paper, but do not guarantee that every request can use the full context or output allowance in a given serving setup.

Open weights and reasoning

Open weights and reasoning support are both listed. Together with text input, text output, and tool support, these properties give developers several ways to evaluate the model for a coding stack. The supplied details do not specify a license, deployment requirements, or a separate price for use beyond the free listing.

Pricing

Free. North Mini Code is listed as free. No paid tier or other price is provided.

Platforms

The supplied details identify a Cohere model and provide its model and Hugging Face identifiers. They do not specify particular SDKs, hosting services, operating systems, or deployment platforms. Teams should verify availability in the platform they plan to use rather than assume support from the model identifiers alone.

Who it's for

North Mini Code is a candidate for developers and teams evaluating a free, open-weight model with an agentic coding focus, text-based interaction, tool support, and reasoning. Its long context and output limits may also be relevant when a workflow needs to provide extensive text or request a lengthy response. Whether it fits a production stack depends on factors not given here, including the required tools, deployment path, and task-specific results.

Teams comparing Cohere models can also consider Cohere models. For discovery by capability, North Mini Code is listed among Free models, Models with tool calling, Open-weight models, Newest models, and Reasoning models.

Pros and cons

Pros

• Free listing with open weights.
• Designed as an agentic coding model and supports tools.
• Reasoning is supported.
• Offers a 256,000-token context window and 64,000-token maximum output.

Cons

• The listed details do not identify supported programming languages, integrations, or deployment platforms.
• No benchmark results or task-specific performance information are provided.
• No license terms or deployment requirements are specified.

Alternatives

For other Cohere options, compare North Mini Code with Command A+, Command A, Command R7B (12-2024), Command R (08-2024), and Command R+ (08-2024). The available information here does not establish how those models compare on coding tasks, pricing, or deployment.

Verdict

North Mini Code brings together an agentic coding focus, tool support, reasoning, open weights, and a free listing. Its context and output limits are clearly stated, but the supplied specifications leave practical fit questions unanswered: supported integrations, deployment options, licensing, and coding performance. It is worth considering when those core capabilities match a team’s requirements, with platform and task fit verified before it becomes part of a working stack.