Claude Code’s default model moves from Sonnet to the newly released Claude Opus 5.5, a 1-million-token model now standard even on Pro and Team Standard plans.
GitHub Copilot rolls out two new OpenAI models, GPT-6 Sol and GPT-6 Luna, joining GPT-6 Astra in its lineup for agentic and interactive coding.
A new global default-enablement policy lets Copilot Business and Enterprise admins decide how unconfigured features behave starting October 22.
Cursor launches Rollouts and Security Review, two automated bots for Teams and Enterprise that watch deployments for regressions and scan PRs for vulnerabilities.
Codex CLI 0.157.0 adds a one-key shortcut to fork a running conversation, fullscreen transcripts by default, and support for GPT-6 Sol and Luna via Amazon Bedrock.
Higgsfield, which aggregates video and image models from Google, ByteDance and Kuaishou into one platform, says it has hit a $1 billion annualized revenue run rate.
Chinese municipalities dangle subsidies and tech-cluster incentives to attract AI video production as local governments race to build an industry around the technology.
Adobe’s September Firefly update adds Kling 3.0 Omni reference-image support for video generation and Boards, plus one-click AI background removal across an entire clip.
Adobe expands its Claude plugin with Acrobat tools for the first time and launches a new Gemini plugin, both giving chat-based AI direct access to Photoshop, Premiere, Illustrator and Lightroom.
DaVinci Resolve 21.1 adds a built-in MCP server so any compatible AI assistant, not just one proprietary integration, can directly execute editing operations inside the app.