According to tech media reports, OpenAI may release its next-generation flagship model, GPT-5.6, in late June. Beyond performance improvements, public interest centers on whether it will further evolve the chat model into an Agent capable of directly executing tasks.
The release time is concentrated next week.
Reports suggest the GPT-5.6 series may include Mini, Standard, and Pro versions, with the most likely release window between June 22 and 28. On the prediction market Polymarket, the probability for this timeframe once rose to about 80%, but later declined amid rumors of a delay.
Some early signs have also raised market expectations. Reports indicated that OpenAI’s Codex backend logs briefly displayed the identifier “gpt-5.6,” and some ChatGPT Pro users reportedly accessed a test version and noted improved output quality under the same prompts.
Execution capability becomes the focus
Based on the information currently available, the most notable aspect of this upgrade is enhanced execution capability. Reports suggest that GPT-5.6’s context window may expand from 1 million tokens to approximately 1.5 million tokens, making it better suited for handling long-form coding tasks and multi-step workflows.
Meanwhile, the model’s token efficiency is reportedly improved by an additional 10% to 15%, enabling the system to process more content at the same cost and making it better suited for enterprise continuous task scenarios.
- Directly convert design mockups into executable code
- Generate a rotatable and scalable SVG 3D object
- Execute browser operations via Playwright
If these capabilities are officially launched, the model's role will no longer be limited to conversation and recommendations but will be able to directly perform actions in certain scenarios.
The low-price strategy may continue to be implemented.
Beyond performance, pricing is also a key focus of these rumors. The report mentions that GPT-5.6’s token pricing could be approximately one-third that of Claude Fable 5, continuing OpenAI’s strategy of competing in the market with lower prices.
However, market expectations for this update are mixed. Some users believe that GPT-5.5’s recent performance fluctuations may be related to backend testing of a new version; others caution that 5.6 is still a minor iteration, and a true model-level leap may not occur until GPT-6.
Overall, if GPT-5.6 is launched on schedule, its impact may extend beyond benchmark scores to whether OpenAI can further productize its flagship model to handle more executable tasks in real-world workflows.