Three regressions confirmed: reasoning effort, context retention, and verbosity changes

Anthropic publishes post-mortem on Claude Code quality issues

Release of a comprehensive cheat sheet for Claude Code aggregating CLI commands, MCP configurations, and workflow shortcuts.

Released the Kimi 2.6 Thinking open-weights reasoning model

Released Playwright MCP Server for generating web tests from natural language prompts

Announced 46:1 context cache ratio for managing 39 projects and integration with FastMCP

Critical security vulnerabilities (CVE-2026-35022) disclosed, requiring immediate patching and cessation of authentication helper use.

Released the open-source coding model Kimi K2.6, achieving top scores on SWE-Bench Pro and HumanEval with Tools benchmarks.

Released a trillion-parameter open-source model matching Claude Opus on coding benchmarks.

Teased upcoming 'Kimi 2.6' code model via leaked image, suggesting a major update to Kimi Chat.