Cohere

North Small Translate

Cohere

North Small Translate is a dedicated text translation model supporting more than 50 languages. Its 218B-parameter mixture-of-experts network activates 25B parameters and targets long-form translation with efficient serving. Open weights are available in several precisions under CC BY-NC 4.0, while commercial use requires appropriate licensing. It can be evaluated through Cohere’s Chat API and a public demo. The release date follows the September 9 API changelog; the launch blog followed on September 10. Translation quality claims depend on the specific language pair, domain, and evaluation method.

Tiny Aya Thinker

Cohere

Tiny Aya Thinker is Cohere Labs’ 3.35B multilingual reasoning research release, published with the Building Multilingual Bridges paper. L2-Thinker produces reasoning in the user’s language across 60 languages, while En-Thinker provides an English-reasoning comparison. The models build on a 32K Tiny Aya base checkpoint. Downloadable research weights and a public demonstration support multilingual reasoning experiments. The paper publication precedes the October 5 explanatory blog; that later blog is not the first release.

North Micro Vision Instruct

Cohere

North Micro Vision Instruct is a compact open-weight vision-language model for captioning, visual questions, grounding, and document understanding. It combines a 2B language model with a 400M vision encoder and preserves image aspect ratios through native-resolution processing. The language backbone supports 128K tokens, but multimodal training and validated operating conditions cover only 8K; this page uses the conservative multimodal limit. Apache-licensed weights make it a foundation for domain-specific fine-tuning and local experimentation. It supports interleaved images and text, multilingual prompts, and text output.

North Mini Code

Cohere

North Mini Code is Cohere’s first North coding model, trained for repository-level software changes and terminal-based agent tasks. Its mixture-of-experts architecture has 30B total parameters and 3B active per token, allowing relatively efficient inference while retaining the memory needs of the full model. Training across multiple agent harnesses targets transferable software engineering behavior rather than a single fixed scaffold. Apache-licensed weights, Chat V2 access, and private Model Vault deployments are available. The documented context limit is 256K with up to 64K output tokens.

Command A+

Cohere

Command A+ unifies Cohere’s enterprise reasoning, retrieval, translation, and image understanding in a sparse mixture-of-experts model. It has 218B total parameters with 25B active per token, supports 48 languages, and combines text and image inputs with text generation. Unlike earlier Command A checkpoints licensed for non-commercial weight use, its downloadable weights use Apache 2.0. Hosted Chat access and private deployment support enterprise agents that must reason over documents and call external tools. Its 128K context and 64K output limits are the documented product limits.

Tiny Aya

Cohere

Tiny Aya is Cohere Labs’ 3.35B multilingual model family designed to make local language generation practical across 70 languages. The family includes a base checkpoint, a globally balanced instruction model, and Earth, Fire, and Water regional variants. These emphasize African and West Asian, South Asian, and European and Asia-Pacific languages respectively. Instruction checkpoints are available through Cohere’s Chat API and downloadable under a non-commercial license. The family’s 8K context and compact size suit offline assistants, translation experiments, and regional-language applications without requiring a large hosted model.

Command A Translate

Cohere

Command A Translate is a specialized 111B-parameter language model for text translation across 23 languages. It adapts the Command A generation to multilingual translation tasks rather than general enterprise agent use. Supported languages span European and Asian languages as well as Arabic, Hebrew, Persian, and Hindi. Cohere provides downloadable research weights under non-commercial terms and a commercial hosted translation model through the Chat API. The documented hosted configuration supports 8K input and 8K maximum output, described as 16K combined in the detailed model documentation. This dedicated translation release is distinct from routine language prompts applied to the standard Command A model.

Command A Reasoning

Cohere

Command A Reasoning is Cohere’s first dedicated reasoning language model, extending the enterprise Command A generation with controllable thinking. Its 111B-parameter dense Transformer supports 23 languages, tool-driven workflows, and document-grounded responses. Applications can switch thinking on or off and set a reasoning budget to trade computation for task performance. The current hosted model provides a 256K context and up to 32K output tokens. Public research weights use non-commercial terms, with commercial inference available separately. Its reasoning specialization is a distinct catalog identity from the standard Command A model and later sparse Command A+ release.

Command A Vision

Cohere

Command A Vision adds enterprise image understanding to a Command A language backbone through a multimodal adapter and SigLIP2 vision encoder. The 112B-parameter model accepts text and images and generates text for OCR, charts, tables, document questions, and visual analysis. It officially targets six languages: English, Portuguese, Italian, French, German, and Spanish. The hosted API documents a 128K context and 8K output; the downloadable Hugging Face configuration defaults to 32K while supporting extension to 128K. Its downloadable weights use non-commercial terms. Cohere’s limitations explicitly exclude tool use despite the capability badges on the documentation page.

Command A

Cohere

Command A is Cohere’s 111B-parameter enterprise language model for retrieval, tool use, agents, and multilingual instruction following. Its dense autoregressive Transformer uses local and global attention to serve long contexts efficiently. Cohere documents 23-language support, a 256K context, and 8K output on its hosted Chat API. The provider reports deployment on two A100 or H100 GPUs, with throughput depending on serving configuration. Downloadable weights are a non-commercial research release, distinct from the separately contracted commercial API. Command A forms the language backbone of later named vision and reasoning releases rather than subsuming them into this page.

Aya Vision

Cohere

Aya Vision is Cohere’s multilingual vision-language family with 8B and 32B language tiers. It combines an autoregressive language model, a SigLIP2 vision encoder, and a multimodal adapter to handle image-and-text prompts in 23 languages. Its capabilities include OCR, image captioning, visual questions, translation of text in images, and document understanding. Both variants support a 16K context and generate text rather than images. Research weights use non-commercial terms. Cohere’s current Chat API serves the 32B checkpoint with a 4K output limit; the 8B hosted checkpoint was retired in April 2026.

Command R7B Arabic

Cohere

Command R7B Arabic is a distinct language specialization for Modern Standard Arabic and English, emphasizing cultural understanding, instruction adherence, length control, and responses in the requested language. Its model card counts approximately 8B total parameters, comprising 7B Transformer parameters and 1B embedding parameters. The optimized autoregressive architecture combines local and global attention and supports a 128K context. Conversational and concise instruction modes use different preambles. It is specifically trained for citation-bearing retrieval generation in Arabic and English. Research weights and a hosted demonstration are available under Cohere’s non-commercial weight terms.

Command R7B

Cohere

Command R7B is a smaller Command model optimized for multilingual enterprise generation, document grounding, code, and multi-step agents. Its optimized decoder alternates sliding-window attention with periodic global attention, combining local efficiency with a 128K context. Cohere positions the model as a 7B tier and releases research weights as well as hosted Chat access. Post-training uses supervised fine-tuning and preference training for instruction following and safety. It supports 23 languages, citation-bearing retrieval templates, and function descriptions supplied as JSON schemas. The current API limits output to 4K tokens.

Aya Expanse

Cohere

Aya Expanse is a multilingual text generation family with 8B and 32B variants, targeting strong instruction following across 23 languages. Its training combines data arbitrage, multilingual preference optimization, and model merging. The 32B model supports a 128K context, while the smaller 8B release supports 8K. Research weights are downloadable under non-commercial terms. Cohere’s current Chat API serves the 32B checkpoint with a 4K output limit; the hosted 8B checkpoint was retired in April 2026. This family page groups the two sizes without treating their context limits or current API access as identical.

Aya 23

Cohere

Aya 23 is an open-weight multilingual language model family with 8B and 35B decoder-only variants. It concentrates on 23 languages to improve depth of language understanding and instruction following compared with Aya 101’s broader encoder-decoder coverage. The models are built on Cohere’s Command generation with multilingual pretraining and instruction post-training, supporting conversation, summarization, translation, and general text generation. Both sizes have an 8K context. Downloadable weights and a technical report support multilingual research, while the non-commercial license governs weight use. This family page groups the two routine size variants.

Command R+

Cohere

Command R+ is the larger 104B-parameter member of Cohere’s Command R generation, designed for complex document retrieval and multi-step tool workflows. It supports grounded responses with citations, multilingual conversation in ten primary languages, and external API or database interactions. The April 2024 release and August 2024 refresh share this family page. Research weights are downloadable under non-commercial terms, while commercial hosted and private deployment options use separate terms. The current API catalog keeps the August checkpoint live and marks the original command-r-plus alias deprecated. Its documented hosted limits are 128K context and 4K output.

Command R

Cohere

Command R is Cohere’s 35B-parameter multilingual language model for retrieval-augmented generation and tool-driven enterprise workflows. It generates answers with document citations and can interact with external services through tool calls. The original March 2024 weights and an August 2024 update are grouped here as releases of the same named model. The family supports a 128K context and ten primary languages, with downloadable research weights and commercial hosted access. The original command-r alias was deprecated in September 2025; the August 2024 checkpoint remains listed as live in the current API catalog.

Aya 101

Cohere

Aya 101 is Cohere’s massively multilingual instruction-tuned research model, built by fine-tuning the 13B mT5 encoder-decoder model. It supports 101 languages, more than half of them lower-resource languages, and uses a broad mixture of instruction examples from the Aya community effort and other datasets. The released model focuses on multilingual generation, question answering, summarization, and translation. Apache 2.0 weights and a published research paper support further study and adaptation. Unlike later decoder-only Aya generations, this model uses an encoder to represent the prompt and a separate decoder to generate the response.

Command

Cohere

Command is Cohere’s original instruction-following language model, first released as the command-xlarge public beta in November 2022. It introduced a generation interface conditioned on direct commands for summarization, rewriting, extraction, and question answering. The family later added the smaller Command Light tier, conversational improvements, and document-grounded Chat access. Its final documented API configuration supported a 4K context. This historical page groups the original Command, Command Light, renamed aliases, and routine nightly updates; it does not duplicate the later Command R or A families. Cohere deprecated the original hosted models in September 2025.