OpenAI Launches GPT-6.1 Sol and Always-On 'Dots' Agents
OpenAI introduced the GPT-6.1 Sol model, always-on 'Dots' agents with dedicated cloud environments, and updated developer tooling.
Key takeaways
- According to OpenAI, GPT-6.1 Sol delivers near-Astra performance on agentic coding, computer use, and professional workflows at one-fifth the token cost.
- OpenAI introduced 'Dots,' always-on autonomous agents powered by GPT-6 Astra that execute workflows 24/7 on dedicated cloud computers.
- Developers can deploy GPT-6.1 Sol via the OpenAI API and Amazon Bedrock, alongside new MCP Extensions and a Sign in with ChatGPT DevKit.
OpenAI has officially launched GPT-6.1 Sol, an upgraded reasoning model designed to provide near-frontier intelligence at a fraction of existing operating costs, alongside Dots, a system of always-on autonomous agents that execute multi-step workflows on dedicated cloud computers. The rollout introduces major price reductions for agentic coding and professional tasks, expanded third-party integrations, and new developer kits.
According to OpenAI's product announcement, GPT-6.1 Sol provides an alternative to GPT-6 Astra, matching or approaching Astra's performance across coding benchmarks, computer-use evaluations, and document processing while pricing input and output tokens at one-fifth the rate of standard Astra usage.
High-Performance Reasoning at Reduced Token Costs
Standard API pricing is set at $2.00 per million input tokens and $10.00 per million output tokens, with cached input pricing dropping to $0.10 per million tokens—a 95 percent discount compared to standard input rates and a 50 percent cut from GPT-6 Sol's cached input pricing.
On software engineering evaluations, OpenAI reports that GPT-6.1 Sol matches GPT-6 Astra on DeepSWE v1.1 at roughly one-fifth the cost per task, exceeding GPT-6 Sol’s benchmark score by 6.4 percentage points at a lower reasoning effort. According to OpenAI, on GDP.pdf, an evaluation testing multi-page document comprehension across legal, financial, and healthcare domains, GPT-6.1 Sol scored higher than Opus 5.5 with fallbacks at less than half the cost per task.
For enterprise workflows, OpenAI states that GPT-6.1 Sol scored 2.2 percentage points above Opus 5.5 on AutomationBench 1.0.6 at medium reasoning effort, while outperforming GPT-6 Sol by 4.8 percentage points. According to OpenAI, on OSWorld 2.0's offline set for computer use, the model scored within 2.1 percentage points of Astra at maximum reasoning effort while operating at approximately one-seventh the cost per task.
Always-On 'Dots' Agents with Dedicated Cloud Environments
Alongside the model release, OpenAI announced Dots, described in OpenAI's introduction to dots as always-on agents powered by GPT-6 Astra. Dots operate 24/7 on their own isolated cloud computers and browsers to handle multi-step tasks across connected applications.
Dots connect to over 4,000 applications through OpenAI's plugin ecosystem and integrate into communication channels including ChatGPT, Slack, and Microsoft Teams. When users are inactive, Dots run background "proactive research" using restricted read-only permissions that prevent unauthorized changes or outbound messaging.
OpenAI is also testing specialist dots for organizations, which receive their own identity for access management, IT-provisioned hardware, and support for deep integrations with corporate systems of record. OpenAI confirmed it is working with Microsoft to integrate specialist dots into Microsoft Agent 365 for enterprise governance.
Dots are rolling out to ChatGPT Pro and Business Premium subscribers in eligible markets, with your first primary dot included at no additional cost. Enterprise, Edu, and Healthcare administrators can enable access through an enterprise beta.
Developer Tooling and Cloud Availability
For software engineers building agentic systems, OpenAI published the openai/mcp-extensions repository, which adds ChatGPT-native entry points to the Model Context Protocol (MCP). The extension library includes support for sidebar entry points, custom file extension viewers, composer mentions, and image picker forms via TypeScript and Python SDKs under an Apache 2.0 license.
OpenAI also published the Sign-in with ChatGPT DevKit, providing local Node.js SDKs and React components for desktop application developers to authenticate users and power local apps using active ChatGPT subscription plans without requiring manual API keys.
Enterprise developers can deploy GPT-6.1 Sol via the OpenAI API as gpt-6.1-sol or through Amazon Web Services. As outlined by Amazon Web Services, GPT-6.1 Sol is generally available on Amazon Bedrock with AWS IAM policy governance, CloudTrail auditing, and zero-operator access isolation on hardware-isolated infrastructure.
Updated ChatGPT Pro Subscription Tiers
To accommodate varying computational requirements, OpenAI revised its ChatGPT Pro offerings, detailed in the OpenAI Help Center. The tier structure now features three monthly plans:
- Pro 100: $100 per month.
- Pro 200: $200 per month, available again for new sign-ups with updated usage allowances. Existing subscribers active as of the cutoff date maintain legacy allowances through October 29, 2026.
- Pro 500: $500 per month, offering the highest included usage allowance and access to Astra Ultrafast.
GPT-6.1 Sol is available immediately in ChatGPT Work, Codex, and via API endpoints, with an Ultrafast speed tier generating tokens up to 8x faster scheduled for release in Codex in the coming days.
Frequently asked questions
What is GPT-6.1 Sol?
According to OpenAI, GPT-6.1 Sol is an upgraded AI reasoning model that delivers performance comparable to GPT-6 Astra on coding, computer-use, and document analysis tasks at one-fifth the token cost.
What are OpenAI Dots?
Dots are always-on AI agents powered by GPT-6 Astra that execute tasks 24/7 on isolated cloud computers, integrating with ChatGPT, Slack, Teams, and over 4,000 third-party apps.
How much does GPT-6.1 Sol cost via API?
GPT-6.1 Sol standard API pricing is $2.00 per million input tokens, $0.10 per million cached input tokens, and $10.00 per million output tokens.
Where can developers access GPT-6.1 Sol?
Developers can access the model via the OpenAI API using the model identifier gpt-6.1-sol, through ChatGPT Work and Codex, and on Amazon Bedrock.
Sources
- Introducing GPT-6.1 SolOpenAI · Official
- Introducing dotsOpenAI · Official
- openai/mcp-extensionsGitHub · Sep 29, 2026 · Official
- openai/sign-in-with-chatgpt-devkitGitHub · Sep 29, 2026 · Official
- About ChatGPT Pro tiers | OpenAI Help CenterOpenAI Help Center · Official
- GPT-6.1 Sol Model | OpenAI APIOpenAI Developers · Official
- Bring near-Astra intelligence to everyday work with GPT-6.1 Sol on Amazon Bedrock | Amazon Web ServicesAmazon Web Services · Sep 29, 2026 · Official
How this story was made: the newsroom picked it up from Google News, theverge.com and 9to5mac.com, gathered the full text of the sources above, and drafted it with AI assistance. Every factual claim was then checked against those sources before publishing (28 claims checked). Illustrations marked as AI-generated are not photographs. Spotted an error? Tell us.
Published September 30, 2026 at 00:10 UTC
