{"id":3075,"date":"2026-09-16T08:27:13","date_gmt":"2026-09-16T08:27:13","guid":{"rendered":"https:\/\/dmarketertayeeb.com\/blog\/north-small-translate-access-deployment\/"},"modified":"2026-09-16T08:27:48","modified_gmt":"2026-09-16T08:27:48","slug":"north-small-translate-access-deployment","status":"publish","type":"post","link":"https:\/\/dmarketertayeeb.com\/blog\/north-small-translate-access-deployment\/","title":{"rendered":"North Small Translate: Access, Languages, Hardware and Cost Boundaries"},"content":{"rendered":"\n\n\n

Short answer:<\/strong> North Small Translate is Cohere\u2019s live, text-only machine-translation model, released on September 10, 2026. Its model ID is north-small-translate-1-0<\/code>; the Hugging Face repository is CohereLabs\/North-Small-Translate-1.0<\/code>. Cohere documents a sparse mixture-of-experts model with 218 billion total parameters and 25 billion active parameters, 16K input and 16K output limits, and support for 50 languages plus documented locale variants.<\/p>\n\n\n\n

There are three different access decisions. You can evaluate it through the Cohere Chat API, where Cohere says it is free until rate limits are reached. You can download the weights from Hugging Face for research and non-commercial use under CC BY-NC 4.0, subject to the Cohere Labs Acceptable Use Policy. For commercial hosted deployment, Cohere documents Model Vault and a commercial license; its current Standard Vault pricing table shows North Small Translate at $17.50 per XL instance-hour, with generative access typically handled through a waitlist. The model is not \u201csmall\u201d in local-hardware terms: Cohere\u2019s minimum examples range from one B200 with NVFP4 W4A16 weights to eight H100s for BF16.<\/p>\n\n\n\n

Cohere\u2019s quality, throughput and cost figures are vendor-reported. The launch blog says the model scored 83.60 across all languages in Cohere\u2019s WMT26 evaluation, but describes GPT-5.6-Sol as the judge. The official WMT26 task describes human evaluation, so the blog figure should not be treated as an official human-evaluated leaderboard result without the exact submission, test set and protocol. A safe first test should measure your own language directions, terminology, formatting, cost, latency and human review load.<\/p>\n\n\n\n

\n\n
Decision point<\/th>Current primary evidence<\/th>Boundary to verify<\/th><\/tr><\/thead>\n
Model identity<\/td>north-small-translate-1-0<\/code>; 218B total \/ 25B active MoE; text in and text out.<\/td>Capture the exact model and checkpoint version used by your account or deployment.<\/td><\/tr>\n
Language scope<\/td>50 languages in the model card; Cohere documentation lists more than 50 language and locale variants.<\/td>Confirm every source-target direction, locale and domain before committing.<\/td><\/tr>\n
API evaluation<\/td>Chat V2 API; free until rate limits are reached.<\/td>North-specific production rate, token price, region and data terms.<\/td><\/tr>\n
Open weights<\/td>Hugging Face download under CC BY-NC 4.0 for non-commercial use.<\/td>Commercial use, derivative terms, acceptable-use obligations and gated repository access.<\/td><\/tr>\n
Hosted production<\/td>Commercial deployment through Cohere Model Vault; Standard Vault table shows $17.50\/XL instance-hour.<\/td>Waitlist, tier, replicas, region, monthly\/annual commitment, encrypted-vault price and support terms.<\/td><\/tr>\n
Local hardware<\/td>Minimum examples: 1 B200 at NVFP4 W4A16 or 2 H100s; BF16 requires 4 B200s or 8 H100s.<\/td>Actual memory headroom, throughput, power, networking, serving version and workload concurrency.<\/td><\/tr>\n<\/tbody>\n<\/table><\/figure>\n\n\n\n

What Cohere released<\/h2>\n\n\n\n

Cohere\u2019s September 10 announcement calls North Small Translate its first translation model in the North family. It is a sparse mixture-of-experts (MoE) model purpose-built for machine translation, rather than a general translation feature hidden inside a broader model. Cohere says the model is available for research and non-commercial use under CC BY-NC 4.0, links to Hugging Face for weights and a hosted demo Space, and points enterprises to RWS\u2019s Language Weaver product for a managed translation and localization offering.<\/p>\n\n\n\n

The current Cohere model documentation marks north-small-translate-1-0<\/code> as Live<\/strong> and exposes it through the Chat endpoint. The model card identifies the research release as North-Small-Translate-1.0<\/code>. Keep those names separate in implementation notes: the Cohere API model ID is hyphenated and lower-case, while the Hugging Face repository uses the Cohere Labs organisation and title-case model name.<\/p>\n\n\n\n

The model is text-only. Cohere documents text input and generated text output, with a 16K input context and a maximum of 16K output tokens. Those limits are translation-call limits, not a promise that every document can be pasted unchanged: markup, instructions and prompt overhead consume context, and long documents still need segmentation and structure-preserving tests.<\/p>\n\n\n\n

Model identity and training lineage<\/h2>\n\n\n\n

The headline \u201cSmall\u201d describes the North family position, not a laptop-sized checkpoint. North Small Translate has 218 billion total parameters and 25 billion active parameters. The Hugging Face model card describes a decoder-only sparse MoE Transformer with 128 experts, eight selected per token and shared experts. It also describes interleaved sliding-window and global attention. These are model-card details, not a guarantee of a particular serving implementation or memory footprint.<\/p>\n\n\n\n

A Cohere-authored research paper presents the system as \u201cCohere CAT+\u201d for the WMT General MT task and says it uses the same foundation as Command A+. The paper describes a non-reasoning base model trained for throughput, followed by a five-step protocol combining coarse and fine-grained supervised fine-tuning, direct preference optimization and online reinforcement learning. It also describes related tasks such as post-editing, quality estimation, structured translation and instruction following.<\/p>\n\n\n\n

That paper explains the research lineage but does not grant a separate fine-tuning recipe to customers. I found no current official North Small Translate adapter, LoRA, fine-tuning API or customer-training workflow. Treat customization as an open question for Cohere rather than assuming that an open-weight download can be fine-tuned, commercially redistributed or served with the same quality and support terms.<\/p>\n\n\n\n

The four practical access paths<\/h2>\n\n\n\n

1. Cohere Chat API for a bounded evaluation<\/h3>\n\n\n\n

Cohere\u2019s North documentation supplies a Chat V2 example using ClientV2<\/code> and the model ID north-small-translate-1-0<\/code>. The docs say North Small Translate is available on the free tier until rate limits are reached. This is useful for a small proof of concept, not an unlimited production grant.<\/p>\n\n\n\n

Cohere\u2019s general rate-limit page distinguishes free trial and paid production keys and says trial keys are limited to 1,000 API calls per month. Its Chat table lists several named models at 20 requests per minute but does not list North Small Translate. Do not assign North the rate of Command A Translate or North Mini Code by analogy. Check the current dashboard, response headers and Cohere support for the exact North limit and any production-key requirement.<\/p>\n\n\n\n

A source-checked but unexecuted example looks like this:<\/p>\n\n\n\n

from cohere import ClientV2\n\nco = ClientV2(api_key=\"<YOUR_API_KEY>\")\nresponse = co.chat(\n    model=\"north-small-translate-1-0\",\n    messages=[{\n        \"role\": \"user\",\n        \"content\": \"Translate everything that follows into French:\\n\\n\"\n                   \"Enterprises need accurate translations of business-critical documents.\"\n    }],\n)\nprint(response.message.content[0].text)<\/code><\/pre>\n\n\n\n
The request pattern follows Cohere\u2019s current model documentation, but it was not executed here because no authorised API credential was available. Before using it in production, add input-size checks, retry\/backoff for rate limits, language-direction validation, output parsing and human review for high-risk content. The example does not prove a key, region, model entitlement or billing state.<\/p>\n\n\n\n
2. Hugging Face weights for research and non-commercial use<\/h3>\n\n\n\n
The Cohere Labs model card<\/a> provides the downloadable open-weight research release and several quantized repositories. Access is gated: the public page says you must agree to the license conditions and share contact information, with collection and sharing governed by the linked Cohere Privacy Policy. A hosted Hugging Face Space lets you try the model before downloading the weights.<\/p>\n\n\n\n
The licence is CC BY-NC 4.0, with a Cohere Labs Acceptable Use Policy addendum. \u201cOpen weights\u201d therefore does not mean \u201ccommercially unrestricted.\u201d A marketing agency, SaaS product or localization vendor should obtain a commercial license before using the weights for paid production work. DMT\u2019s open-weight licensing analysis<\/a> provides broader context on why a downloadable checkpoint and a commercial right to use it are separate questions. The Hugging Face page also reports that the base repository is not deployed by an Inference Provider, so a one-click hosted provider route should not be assumed.<\/p>\n\n\n\n
3. Model Vault for commercial hosted inference<\/h3>\n\n\n\n
Cohere documents commercial North deployment through Model Vault. A Standard Vault is a Cohere-managed, single-tenant inference environment: Cohere manages serving infrastructure, maintenance, updates, scaling and monitoring while the customer calls a vault endpoint. The North integration can run the North application and control plane in the customer\u2019s VPC while routing model inference to the vault.<\/p>\n\n\n\n
Standard Vault pricing currently lists North Small Translate at $17.50 per XL instance-hour<\/strong>. The table shows no L-tier rate for this model. Generative models typically require a waitlist, and Cohere says monthly and annual commitment pricing follows fixed or flex plans and requires contacting Cohere. The hourly figure is a Model Vault instance rate, not a complete translation cost: add the replica range, uptime, autoscaling, storage, network, application, human review and license terms before comparing it with a token-priced API.<\/p>\n\n\n\n
Model Vault\u2019s standard deployment is protected in transit and at rest and is logically isolated. If Zero Data Retention (ZDR) is enabled for a Standard Vault standalone deployment, Cohere says it processes inputs and outputs for inference without retaining prompts or responses. Encrypted Vault adds hardware-backed confidential computing and remote attestation, but model and GPU availability differs; verify that North Small Translate is offered in the chosen vault type, tier and region.<\/p>\n\n\n\n
4. Private VPC or on-premises deployment<\/h3>\n\n\n\n
Cohere\u2019s deployment-options material describes private cloud, VPC and on-premises paths for North and says private deployments keep prompts, outputs and fine-tuned models in the customer environment, with Cohere unable to access that data or infrastructure. That is an enterprise deployment option, not an automatic entitlement for every North model. Confirm that North Small Translate, the desired quantization and the selected serving stack are supported for the customer\u2019s cloud, hardware, region, license and support plan.<\/p>\n\n\n\n
Language coverage: 50 languages is not a pair-by-pair guarantee<\/h2>\n\n\n\n
The model card lists 50 languages: Albanian, Arabic, Bengali, Bulgarian, Catalan, Chinese (Simplified), Chinese (Traditional), Croatian, Czech, Danish, Dutch, English, Estonian, Filipino, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Irish, Italian, Japanese, Korean, Latvian, Lithuanian, Malay, Maltese, Norwegian, Persian, Polish, Portuguese, Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Swedish, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu and Vietnamese.<\/p>\n\n\n\n
Cohere\u2019s current documentation describes more than 50 language and locale variants and calls Modern Standard Arabic, German, French, Japanese, Korean, Russian and Ukrainian its tier-one languages. It separately lists variants such as Egyptian, Modern Standard and Saudi Arabic; Portuguese (Portugal); Norwegian Bokm\u00e5l; and Serbian (Cyrillic). The launch blog uses \u201c50+ languages,\u201d while the model card uses 50, so use the full current documentation list rather than repeating an unqualified count.<\/p>\n\n\n\n
Neither the launch blog nor model card supplies a complete source-target matrix with quality or availability per direction. \u201cSupports 50 languages\u201d should therefore mean \u201cthe model is documented for those languages,\u201d not \u201cevery one of the 2,500 possible directions has equal quality, locale support or commercial availability.\u201d Test the directions that matter to your business, including regional variants, terminology and script handling.<\/p>\n\n\n\n
Hardware, quantization and serving boundaries<\/h2>\n\n\n\n
\n\n
Checkpoint format<\/th>Cohere model-card example<\/th>What it means for planning<\/th><\/tr><\/thead>\n
BF16<\/td>4 \u00d7 B200 or 8 \u00d7 H100<\/td>Highest listed precision; large multi-GPU footprint.<\/td><\/tr>\n
FP8<\/td>2 \u00d7 B200 or 4 \u00d7 H100<\/td>Lower-precision checkpoint; confirm serving support and quality on your language pairs.<\/td><\/tr>\n
NVFP4 W4A16<\/td>1 \u00d7 B200 or 2 \u00d7 H100<\/td>Smallest listed hardware example; it is still a datacenter-GPU deployment.<\/td><\/tr>\n<\/tbody>\n<\/table><\/figure>\n\n\n\n
Cohere calls these example minimum requirements. They are not a throughput or total-cost guarantee. A 25B active-parameter MoE can reduce per-token computation, but the 218B total parameter set, temporary buffers, KV cache, networking and concurrency still affect memory and cost. Do not turn \u201cactive parameters\u201d into a claim that the model fits in 25GB or on a 25B-class card.<\/p>\n\n\n\n
The model card recommends installing a Transformers version from the source repository containing the required model changes and recommends greedy decoding, which Cohere says its production deployment uses. It warns that the older translation pipeline is no longer supported in Transformers 5; loading the model directly or using a compatible earlier Transformers version is required. Generated replies contain structural markers, and Cohere recommends parsing them with its cohere_melody<\/code> library. The official vLLM example uses tensor parallelism and Cohere\u2019s parser settings, but the example is not a substitute for a compatibility test on the chosen checkpoint.<\/p>\n\n\n\n
The following deployment questions remain open in the reviewed sources:<\/p>\n\n\n\n