September is a big month for Big Tech events, where companies showcase their latest products. And now, OpenAI has just joined the party with DevDay 2026 in San Francisco.
The AI giant held its annual DevDay event on Tuesday and had a slew of announcements to make. OpenAI recently released GPT-6 Astra, an advanced new model, but it’s also facing demands for much stricter regulation after its AI agents hacked third-party systems.
The biggest news of the day was the launch of Dots, a new AI agent to rival Meta’s Muse AI.
Here are all the biggest announcements from OpenAI DevDay 2026.
Dots
Credit: OpenAI
Apparently, every AI company needs cartoon mascots nowadays.
OpenAI is taking a page out of Meta’s book, and it’s not just the animated characters. Dots are OpenAI’s always-on AI personal assistant product, a la Meta’s Muse.
According to OpenAI, Dots work and complete tasks across existing context and connected applications. As users work with their Dots assistant, it will learn from user feedback and improve on said tasks over time. Being an always-on AI agent, Dots work in the background via a secure cloud computer as users go about their day, OpenAI said.
For now, Dots are available to users on the Pro and Business Premium enterprise plans. The new AI agents began rolling out to eligible users today.
Mashable Light Speed
GPT-6.1 Sol
OpenAI also announced an upgraded version of GPT-6 Sol, the model that can be viewed as a more cost-efficient version of GPT-6 Astra, OpenAI’s most powerful model.
OpenAI says GPT-6.1 Sol has “exceptionally strong performance on agentic coding, computer use and professional work.” It’s nearly as intelligence as Astra but costs less than a quarter the price.
Here’s how the pricing breaks down:
-
GPT 6.1 Sol costs $2 per million input tokens
-
$0.10 per million cached input tokens
-
$10 per million output tokens
Pro 500 plan and Ultrafast
If you thought OpenAI’s $200 per month plan was expensive, wait until you hear about Ultrafast.
OpenAI announced a new $500 per month plan called Pro 500. The new plan includes access to Ultrafast, a feature that allows OpenAI’s Codex and Work apps to work on tasks faster. Ultrafast allows Codex and Work to generate up to 300 tokens per second, or 8 times faster.
“Ultrafast is our premium speed tier for workloads where speed matters most,” OpenAI said.
The feature is also coming to the API, which will provide 6 times the standard speed.
ChatGPT Space
AI collaboration will now happen in a new dedicated space, aptly called ChatGPT Space. Colleagues can work together on ChatGPT Space, sharing context, and their Dot AI assistants can access this shared knowledge.
More of OpenAI’s DevDay announcements can be found at the DevDay 2026 website.
Disclosure: Ziff Davis, Mashable’s parent company, in April 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.
Topics
Artificial Intelligence
OpenAI
