Cohere

Volume 342, Issue 3 · 7 articles from 3 issues

In This Issue

• Automation's Early Footprint: The ATE Dataset
• How Small Models Can Make a Big Impact for Enterprises
• Introducing North Small Translate: A Leading Sovereign Open-Weight Machine Translation Model
• Inside the Megakernel Serving Engine for North Mini Code
• Machine Translation Is Still Broken for Most of the World's Languages: Cohere Builds Non-Reasoning for a Reason
• Cohere's Model Vault Now Encrypts AI Inference so Even Cohere Cannot See Enterprise Customers' Data
• Who Gets to Define the Rules for AI?

Automation's Early Footprint: The ATE Dataset (Cohere, Thursday, September 3rd, 2026)

Cohere studied 696,000 AI tools and found only 2.6% truly automate work, releasing the Agentic Task Ecosystem dataset. Read more

How Small Models Can Make a Big Impact for Enterprises (Cohere, Wednesday, September 2nd, 2026)

Cohere explains where small language models beat frontier models for enterprises and how to build a small model strategy. Read more

Introducing North Small Translate: A Leading Sovereign Open-Weight Machine Translation Model (Cohere, Thursday, September 10th, 2026)

A 218B-parameter translation model covering 50-plus languages, released open weights under a non-commercial licence. Read more

Inside the Megakernel Serving Engine for North Mini Code (Cohere, Tuesday, September 8th, 2026)

A technical account of the megakernel approach delivering 1.58x faster LLM serving on H100 devices. Read more

Machine Translation Is Still Broken for Most of the World's Languages: Cohere Builds Non-Reasoning for a Reason (The New Stack, Sunday, September 13th, 2026)

The New Stack examines Cohere's North Small Translate, a 218B-parameter MoE open-weight translation model built deliberately non-reasoning. Read more

Cohere's Model Vault Now Encrypts AI Inference so Even Cohere Cannot See Enterprise Customers' Data (VentureBeat, Wednesday, September 16th, 2026)

VentureBeat reports that Cohere's Model Vault adds confidential-computing inference that hides customer data from Cohere itself. Read more

Who Gets to Define the Rules for AI? (Cohere, Sunday, September 13th, 2026)

Cohere argues AI standards should rest on evidence rather than being set by a small group of incumbent labs. Read more

More from Cohere · All section posts · This week