🚀 OpenAI Launches GPT-6.1 Sol

PLUS : OpenAI Dots

Author

The Summary AI
September 30, 2026

Sponsored by

Welcome back!

OpenAI just replaced GPT-6 Sol after only seven days. The new GPT-6.1 Sol seemingly responds to the launch of Claude Opus 5.5, getting close to Astra-level performance at one-fifth of the cost. Let’s unpack…

Today’s Summary:

  • 🚀 OpenAI launches GPT-6.1 Sol

  • 🔥 OpenAI launches persistent Dots agents

  • ⚡ Claude Sonnet 5.5 nears Opus

  • 🛡️ OpenAI delays Astra 6.1 over safety

  • 🧬 Claude discovers new enzyme system

  • 🥇 Xiaomi MiMo becomes top open model

  • 🛠️ 2 new tools

TOP STORY

OpenAI launches GPT-6.1 Sol

The Summary: OpenAI has launched GPT-6.1 Sol, following a mixed response to the recent GPT-6 Sol and comparisons with Anthropic’s Opus 5.5. The new model comes close to GPT-6 Astra on coding, computer use, and professional work at about one-fifth of Astra’s cost.

Key details:

  • Artificial Analysis puts Sol one point below Astra, while costing only $0.72 per task versus $3.26. Cached input falls to $0.10/M tokens, making agent use far cheaper

  • Available today in ChatGPT Work and Codex for Plus, Pro, Business, and through the API; not available on regular Chat

  • Codex now runs in the cloud. Persistent environments keep tasks running after you close your laptop

  • Ultrafast mode pushes Astra to 300 tokens/sec, at 6x the price

  • OpenAI added a $500 Pro plan with 25x Plus usage. The $200 Pro plan drops from 20x to 10x usage after October 29

Why it matters: OpenAI now owns the floor of the frontier: there is no cheaper model at the intelligence level of GPT-6.1 Sol. At the same time, Claude Opus 5.5 holds the ceiling of the frontier, where users prefer to pay a premium to get the maximum intelligence for coding and hard tasks.

FROM OUR PARTNERS

Your AI Agent Follows You Everywhere

AI agents built to get work done

Skydive agents take on work for you, using the same tools your team already uses.

Talk to your agent on the web, Slack, email, iMessage, or right from your terminal. Wherever you pick up the conversation, your agent keeps the context.

Put agents to work across customer support, sales, marketing, engineering, ops, and more.

Give your agent a job. They’ll take it from there.

OPENAI

OpenAI launches Dots, personal AI agents

The Summary: OpenAI has launched Dots, personal agents that keep working after you close ChatGPT. Each Dot runs on GPT-6 Astra, gets its own cloud computer and browser, and reaches more than 4,000 apps through plugins. It watches projects, hands work to Codex, and pings you in Slack or Teams when it needs a decision.

Key details:

  • A Dot can run several projects at once on its own cloud computer

  • When idle, Dots can continue scanning connected apps to prepare useful work in a read-only proactive mode

  • Multiple specialist Dots can get their own identity and credentials

  • Rolling out to Pro and Business Premium users. Delayed for regulatory approvals in the EU, Switzerland, and the UK.

Why it matters: Dots is OpenAI's answer to Meta’s Muse, xAI Grok Bot, and OpenClaw, and it primarily aims at the office. It comes with Pro and Business plans, lives in Slack and Teams, and specialist Dots can get IT-issued credentials. Muse is for end consumers, Grok Bot for developers, and OpenClaw for tinkerers.

FROM OUR PARTNERS

Why Most AI Tools Fail in the Real World

New Harvard Business Review Analytic Services report: why most AI tools fail at B2B customer service, and what to ask before you invest. Get the free report.

ANTHROPIC

Claude Sonnet 5.5 lands two points behind Opus 5.5

The Summary: Anthropic released Claude Sonnet 5.5, a mid-tier model that lands just two points behind Opus 5.5 on the Artificial Analysis Intelligence Index. It runs 30% faster than Sonnet 5 with the same $2/M input and $10/M output price. It targets everyday work like documents, slides, coding and UX.

Key details:

  • Scored 70.6% on Terminal-Bench 4.0, up from 10.3% for Sonnet 5

  • Its sweet spot is Medium effort. Avoid Max effort, which overthinks and burns too many output tokens

  • Live in Claude chat and the API. New default model on the Claude free tier

Why it matters: Sonnet 5.5 gives Claude users a cheap coding daily driver to complement Opus 5.5. At Low and Medium effort it beats Sonnet 5’s best score for about a tenth of the cost, and it still lands very close to Opus 5.5. GPT-6.1 Sol is still cheaper, but it runs only in Work, Codex and the API, while Sonnet can also be used in chat.

TOOLS

🥇 New tools

That’s all for today!

If you liked the newsletter, share it with your friends and colleagues by sending them this link: https://thesummary.ai/