GPT-6.1 Sol from OpenAI is now available in GitHub Copilot and Microsoft Foundry.
In early testing, it completed tasks with fewer tokens and steps than earlier models in the GPT-6 and GPT-5.6 families, making it more affordable to build and run agents at scale.
When an agent’s guardrail says “deny,” the action needs to stop.
Agent Hooks gives builders a shared control contract and conformance tests to verify enforcement across frameworks. Explore the spec, five SDKs, and no-key demos.
Introducing run-assert-eval.
With a single prompt in @code, the skill identifies risks specific to your agent, measures how often they occur, generates runtime policy based on those findings, and reruns the same eval to see whether the policy worked.
Claude Opus 5.5, Anthropic’s newest Opus model, is now available in GitHub Copilot and Microsoft Foundry.
In early testing, Opus 5.5 resolved tasks comparably to Claude Opus 5 while using significantly fewer steps and tokens.
