Hey Devs! This week, I picked three updates that can change your everyday AI workflow, plus a new video and a practical tool to explore.
GPT-6.1 Sol arrives
OpenAI’s new model launched on September 29 and is available in the API, Codex, and ChatGPT Work, with standard API prices of $2 per million input tokens and $10 per million output tokens. Try it on one real bug and compare the result, review time, and cost with your current model.
Read more: https://openai.com/index/introducing-gpt-6-1-sol/
Claude Sonnet 5.5 speeds up everyday work
Anthropic’s September 28 release generates output more than 30% faster in its tests, with unchanged token prices and claimed savings from using fewer tokens. If your integration has thinking disabled, check the required between_tools migration before upgrading.
Read more: https://www.anthropic.com/claude-sonnet-5-5
Copilot gets repeatable agent workflows
GitHub’s October 1 public preview lets you define steps in code, run agents in parallel, and pause for human review. It’s available on all Copilot plans, with experimental features required in the CLI; start with a release checklist and keep deployment approval manual.
Read more: https://github.blog/changelog/2026-10-01-dynamic-workflows-in-copilot-cli-and-the-copilot-app/
From my work
Watch: My 4-minute OpenAI Dots explainer video.
Try: My Salary After Tax Calculator compares estimated take-home pay across 20 countries. Switch currencies and review the country-specific assumptions; results use a standard employee profile and aren’t personal tax advice.
What’s one repetitive development task you’d like an AI agent to handle? Leave an example in the comments.
Cheers, proflead! ;)