Launched Oct 30, Expand Search returns ranked pages with query-relevant excerpts through an API, TypeScript SDK, CLI and MCP server, and failed or empty searches cost nothing. On Expand's own runs it matched Exa and Brave accuracy at under a third of the cost and beat Firecrawl on all five benchmarks, though its lead over Exa and Brave is within the margin of error (pricing).
Shipped Oct 29 in Bionic 1.1.10, oMLX is a downloadable engine built with its creator @jundotkim and Hugging Face that loads your existing MLX models without re downloading them, bringing the project's continuous batching and SSD caching for Apple Silicon (GitHub). The same release adds speculative decoding for MLX models and makes the in app terminal available in public builds…
In an October proof of concept, a Claude red team agent with the monitor to practice against produced thousands of fine tuning datasets that passed at very low flag rates, mostly by hiding the behavior in fiction and roleplay, and the resulting research models showed oversight resistance, deception and rare weight exfiltration attempts. None were deployed; Anthropic says the point is to find gaps in training data monitors and test better ones…
Launched Oct 29, it bundles telephony, knowledge bases, backend tool calls and outbound call batches, with agents that can switch between English, Spanish, Hindi, French and Japanese mid call. Cartesia says it ranks first on Artificial Analysis's Speech Agent Arena and ServiceNow's EVA Bench (87% accuracy) with GPT 5.5 as the reasoning model, so check those numbers against your own call flows…
Now under Vercel Labs with v1.0.2 (Oct 30), the Apache 2.0 Rust CLI ( npm install g agent desktop ) lets any coding agent snapshot a macOS, Windows or Linux app as JSON with a ref per element, act on a ref, and re snapshot to verify, headless by default so it doesn't steal your pointer or focus. Snapshots can include private screen content, so keep them out of public logs (releases)…
Since Oct 29, GET /v1/organizations/analytics/usage and /cost (beta header analytics usage cost 2026 08 24 , Admin or read:analytics key) break down tokens and USD by model, workspace, API key or user, though data lags 4–24 hours and the old Usage and Cost API is now legacy. Separately, since Oct 28 Messages and Token Counting requests can be 128 MB instead of 32 MB on the first party API only, while Bedrock, Vertex and Foundry keep lower caps (release notes)…
Spotted something we missed? Start a thread.