GPT‑6.1 Sol: a new model for coding and agents
What the launch changes and how teams can plan a small, measurable pilot.
The short answer
OpenAI introduced GPT‑6.1 Sol on 29 September. This is announcement analysis; we have not run a hands-on model test.
Announcement date: 29 Sept 2026
What changed?
OpenAI positions Sol for complex coding and agent tasks. Its ChatGPT launch targets Work and Codex; access depends on the plan and, in organizations, administrator settings.
Who is this relevant to?
Teams with recurring development tasks can start with a bug report, a change to an existing codebase and a clearly defined research task.
What this means for your work
Our recommendation: Keep previous outputs as a reference. Give both versions the same requirements and assess working results, necessary corrections and total completion time. A polished explanation alone does not make a software fix successful.
Limits of this report
Provider positioning is not a measurement of your project. Reasoning settings, tools and the testing environment also affect results. Record those factors before changing an established workflow.
Videos to explore
Original provider and creator videos, in English. Titles belong to their authors; their assessments are not Findbest test results. Preview artwork consists of Findbest illustrations.
OpenAI DevDay 2026 Keynote (FULL)
OpenAI · English
A connection to YouTube is made only when you load the video. Your IP address and technical device information will be transmitted.
Watch on YouTubeRead next
AI research: from answer to reliable source
A compact workflow for checking claims transparently.
Introduce AI to a small team: run a pilot before adding subscriptions
Use real tasks, a simple evaluation sheet and a clear decision to find out whether an AI tool helps your team.
Check AI answers: turn source links into supported claims
A source link is not enough. Check what it actually supports and whether the context matches.