Models
Salesforce Koa Is a CRM Reasoning Model Built on NVIDIA Nemotron
Salesforce Koa is a CRM reasoning model built on NVIDIA Nemotron 3 Super. It is in a select pilot, with US availability expected in winter 2026.
Read more
Topic
Models
Frontier models, comparisons, and how to choose what fits your work.
47 articles
Models
IBM Separates Shallow Quantum Circuits and Restricted LLMs on Theoretical Tasks
IBM researchers proved theoretical separations between shallow quantum circuits and restricted LLMs. Here is what the result does and does not show.
Read more
Models
Claude API Adds Background Conversation Compaction for Long-Running Agents
Claude API compaction replaces old message history with a signed summary block while preserving a recent tail. This guide maps the new failure modes.
Read more
Models
Google Launches Gemini 3.8 Live With Extended Thinking for Voice Agents
Gemini 3.8 Live adds a production audio model and an extended-thinking variant. Here are the migration changes, state signals, limits and prices.
Read more
Models
Prompt, RAG, Fine-Tune or Train? A Practical AI Customization Guide
AWS published an eight-step model customization spectrum. This vendor-neutral guide maps knowledge, behavior, latency and cost failures to the right step.
Read more
More stories
NVIDIA Reports a 10.4x JAX MoE Training Gain on GB200
Cohere North Small Translate Has 218B Parameters, but Its Open Weights Are Non-Commercial
IBM Granite Time Series Ships a 385M-Parameter Zero-Shot Forecasting Model
AWS Targets Two LLM Latency Problems With Two Different Caches
AWS Says GPT-5.6 Luna Beats GPT-5.4 Mini on Cost per Correct Answer
Cognition SWE-2 Reaches Devin With a 64% Cost Claim and One Benchmark Warning
Gemini 3.8 Flash Cyber posts 47.2% on CWE-bench. Fairwind controls access
Gemini 3.8 Flash keeps 3.7 pricing. Your completed-task cost can still rise
Claude Fable 5.1 made protein binders and mapped Venus. What the evidence proves
Flower Endeavor 1.0 posts frontier scores. Five deployment facts are still missing