OpenAI releases GPT-6.1 Sol for Codex users, 20% cheaper with Astra-like performance
OpenAI introduced GPT-6.1 Sol as the default model in Codex on September 29, positioning it as a cost-effective alternative to GPT-6 Astra for professional tasks. Available to Plus, Pro, Business, Enterprise, and Edu users (with admin approval for the latter two), the model delivers performance nearly equivalent to Astra at one-fifth the cost, according to OpenAI’s own benchmarks. Users report…
Key points
- GPT-6.1 Sol replaces GPT-6 Sol as Codex’s default model, offering Astra-like performance at 20% lower API costs
- Users report 30% fewer tokens used and faster task completion, but Plus quota limits show no measurable difference for small tasks
- OpenAI warns of higher task-completion persistence (23.5% of attempts) and advises strict permission settings for delegated work
The model handles 1.05M tokens of context and supports five reasoning strengths (low to max), though deeper reasoning increases cost and time. OpenAI warns that while GPT-6.1 Sol is less error-prone (7.7% factual errors at low settings vs. 11.4% for GPT-6 Sol), its persistent task completion rate (23.5%) raises risks for delegated work—users must explicitly set permission boundaries to avoid unintended actions. New features like cloud execution (Plus+) and GitHub integration for code review are included, but Security Cloud remains Pro-tier only.
Model pages: GPT-6.1 Sol → · GPT-6 Sol → · GPT-6 Astra →
The story so far
7 episodes →- OpenAI releases GPT-6.1 Sol for Codex users, 20% cheaper with Astra-like performancethis story
Evaluating Codex's New Model GPT-6.1 Sol for Professional Use: Can Plus Users Access It? Is It Smarter? Does It Use Less Quota? We Tested It
note.com · 2 October 2026
Loading the full article…
This text was published by note.com and written by ママゴトラボ. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Agents & Tools
All →- Microsoft releases Agent 365 to manage AI agents list · 1 src
- Wagtail team reports 50% failure rate in one-month GLM 5.3 Flash challenge · 1 src
- Google Antigravity lets users delegate tasks to AI agents with files and folders · 1 src
- Cloudflare launches Clef decision models to compete with TypeSafe’s Jev · 1 src
- AWS shows multi-turn RL fine-tuning for search agents on SageMaker AI · 1 src
Comments
via GitHub Discussions