OpenAI announced a suite of new products at Dev Day 2026 in San Francisco on Tuesday, headlined by GPT-6.1 Sol, an upgraded language model positioned between the company’s cost-efficient Sol tier and flagship Astra, according to Mashable. The upgraded model delivers “exceptionally strong performance on agentic coding, computer use and professional work” whilst maintaining the $2 per million input tokens pricing of its predecessor.
Key takeaways
- GPT-6.1 Sol costs $2 per million input tokens and $10 per million output tokens, matching the original GPT-6 Sol pricing whilst delivering near-Astra intelligence
- Dots, OpenAI’s new always-on AI agent platform, launched for Pro and Business Premium users with access to 4,000+ app integrations
- The new Pro 500 plan costs $500 per month and includes Ultrafast mode, generating up to 300 tokens per second
- ChatGPT Space enables team collaboration with shared context and Dot AI assistants
- Dots are unavailable in the European Economic Area, Switzerland, and the United Kingdom due to regulatory constraints
GPT-6.1 Sol performance and positioning
OpenAI characterised GPT-6.1 Sol as delivering intelligence nearly matching GPT-6 Astra whilst costing “less than a quarter the price”, as reported by Mashable. The upgraded model maintains the 1.05-million-token context window established across OpenAI’s GPT-6 family and targets agentic workflows—autonomous AI systems that plan, execute, and adapt across multiple steps without constant human oversight.
According to OpenAI’s announcement, GPT-6.1 Sol excels at agentic coding, computer use, and professional work. The model sits between GPT-6 Sol and GPT-6 Astra in OpenAI’s hierarchy, offering developers a middle tier for applications requiring sophisticated reasoning without Astra’s premium cost structure.
GPT-6.1 Sol pricing breakdown
The pricing structure for GPT-6.1 Sol represents no increase over the original Sol tier despite the performance improvements. OpenAI set rates at $2 per million input tokens, $0.10 per million cached input tokens, and $10 per million output tokens, according to the Dev Day announcement covered by Mashable.
| Model | Input ($/1M tokens) | Output ($/1M tokens) | Context window | Position |
|---|---|---|---|---|
| GPT-6 Luna | $0.10 | $0.50 | 1.05M | Entry tier |
| GPT-6 Sol | $2.00 | $10.00 | 1.05M | Mid tier |
| GPT-6.1 Sol | $2.00 | $10.00 | 1.05M | Upgraded mid tier |
| GPT-6 Astra | $10.00 | $50.00 | 1.05M | Flagship |
Developers evaluating cost versus capability can use the AI API cost calculator to model expenses across OpenAI’s GPT-6 family and competing models from Google, Anthropic, and other providers.
Dots always-on AI agents launch
OpenAI introduced Dots, a new always-on AI agent platform positioned as a rival to Meta’s Muse AI, which recently climbed app store charts. Dots are personal assistant agents that “work and complete tasks across existing context and connected applications”, according to Mashable’s coverage of the Dev Day announcements.
Each Dot operates on its own private, secure cloud computer and runs continuously in the background as users go about their day. The agents learn from user feedback over time and can handle tasks 24/7, OpenAI stated. Through an ecosystem of plugins, Dots can connect to more than 4,000 applications, giving them tools to assist across a wide range of workflows.
The launch follows Apple’s recent Siri AI app release and Meta’s Muse AI, reflecting an industry-wide push towards agentic AI in 2026. As Mashable noted, AI industry leaders have recently articulated a vision in which everyone will have personal AI agents handling tasks across apps, devices, and data. However, a group of large banks warned last week that as customers grant AI agents access to devices and financial information, there is growing risk of AI-enabled scams and fraud.
Dots availability and guardrails
Dots began rolling out to eligible users on Tuesday, but availability is limited to OpenAI’s $100-per-month Pro plan and Business Premium enterprise customers, according to the announcement. Each customer on these plans receives access to their first Dot at no extra cost. The agents are not available in the European Economic Area, Switzerland, or the United Kingdom—geographic restrictions that mirror those facing Apple’s Siri AI due to stricter AI regulations in Europe.
OpenAI stated that guardrails are built into Dots to protect users from “harmful behaviour”. A monitoring system watches for safety concerns and can pause a Dot’s work. Before a Dot takes sensitive actions such as changing passwords, OpenAI’s auto-review system checks in with the user. Companies can also create custom dots tailored to their organisation’s specific needs.
Users can open their Dot’s computer at any time to inspect its work. Dots can connect to other devices beyond their own cloud computer, and users can grant permission for a Dot to connect to and use their laptop, where it can then work directly alongside them.
Pro 500 plan and Ultrafast mode
OpenAI announced a new $500-per-month subscription tier called Pro 500, five times the cost of the existing $100 Pro plan. The premium tier includes access to Ultrafast, a feature that accelerates OpenAI’s Codex and Work applications, according to Mashable.
Ultrafast allows Codex and Work to generate up to 300 tokens per second—eight times faster than standard speed. “Ultrafast is our premium speed tier for workloads where speed matters most,” OpenAI stated in its announcement. The feature is also coming to the API, where it will provide six times the standard speed for developers building AI coding agents and other latency-sensitive applications.
ChatGPT Space for team collaboration
OpenAI unveiled ChatGPT Space, a dedicated environment for AI-powered team collaboration. Colleagues can work together in ChatGPT Space, sharing context, and their Dot AI assistants can access this shared knowledge, enabling coordination between human team members and their respective AI agents.
The collaborative workspace extends OpenAI’s product line beyond individual use cases into enterprise team environments, where shared context and knowledge management become critical for productivity.
Frequently asked questions
How much does GPT-6.1 Sol cost compared to GPT-6 Astra? GPT-6.1 Sol costs $2 per million input tokens and $10 per million output tokens, whilst GPT-6 Astra costs $10 per million input tokens and $50 per million output tokens. OpenAI stated that GPT-6.1 Sol costs less than a quarter the price of Astra whilst delivering near-equivalent intelligence.
Who can access Dots AI agents? Dots are available to users on OpenAI’s Pro plan ($100 per month) and Business Premium enterprise customers. Each eligible customer receives access to their first Dot at no extra cost. Dots are not available in the European Economic Area, Switzerland, or the United Kingdom.
What is the Ultrafast feature in the Pro 500 plan? Ultrafast is a premium speed tier included in OpenAI’s $500-per-month Pro 500 plan. It allows Codex and Work applications to generate up to 300 tokens per second, eight times faster than standard speed. The feature is also coming to the API with six times standard speed.
How do Dots differ from Meta’s Muse AI? Both Dots and Muse AI are always-on personal assistant agents that run on private cloud computers. Dots are powered by GPT-6 Astra, can connect to over 4,000 apps through plugins, and learn from user feedback over time. Like Muse, Dots feature cartoon mascots and operate continuously in the background.
What tasks can GPT-6.1 Sol handle? According to OpenAI, GPT-6.1 Sol has “exceptionally strong performance on agentic coding, computer use and professional work”. The model is designed for autonomous workflows that require planning, execution, and adaptation across multiple steps.
The bottom line
OpenAI’s Dev Day 2026 announcements position the company deeper into agentic AI and enterprise workflows. GPT-6.1 Sol offers a capability upgrade at the same pricing as its predecessor, appealing to developers who need near-flagship performance without Astra’s premium cost. Dots represent OpenAI’s entry into the always-on AI agent market alongside Meta and Apple, though geographic restrictions and security concerns around AI agent access remain unresolved industry-wide. The $500 Pro 500 plan targets users for whom generation speed justifies premium pricing—a bet that latency-sensitive workflows will sustain five-figure annual subscriptions. More than 20 announcements were made at the event; full details are available at OpenAI’s Dev Day 2026 website.
Sources: mashable.com. Reported September 30, 2026.