Home
Codex Provides One Million New Contextual Data Points for GPT-5.6 Sol, Available to ChatGPT Users Too
Engineer Tibo, who oversees Codex and ChatGPT at OpenAI, shared news on X indicating that the 1 million-token context window feature of the GPT-5.6 Sol model within Codex is now accessible beyond API key holders and is also available to users of ChatGPT accounts. This change eliminates the previous limitation that forced subscribed users to cut off long-term project histories when dealing with large-scale tasks.

Subscribed users can manually activate the million-token context setting
Tibo explained that GPT-5.6 Sol in Codex now supports a context length of 1 million tokens, a capability previously reserved exclusively for API key holders. He noted that the feature has already been activated through ChatGPT accounts and that users with Plus, Pro, or other subscription plans can manually configure it to utilize the full 1.05 million-token capacity when working with the Codex programming agent.
A larger context window enables Codex to retain more code snippets, tool outputs, and conversation history before automatic compression kicks in, which is particularly beneficial for professional use cases such as large-scale codebase restructuring, intricate debugging, and extended sessions. For developers managing production-level repositories, this improvement directly addresses the previous issue of frequent context exhaustion.
The default setting was carefully chosen; enabling it requires careful consideration
Tibo also stressed that the current default context length represents the outcome of balancing performance needs with cost considerations rather than being set arbitrarily. He advised users to make an informed decision based on their own requirements. Several developers added that, thanks to Codex’s server-side compression technology, the default configuration is sufficient for most routine tasks.
Nevertheless, some users have reported that activating the million-token context setting leads to faster consumption of their subscription quotas, with a single intensive session potentially using up a significant portion of their weekly allowance. Experts recommend enabling this feature only when necessary rather than as the default setting for daily use, to prevent hitting usage limits prematurely on critical projects.
Related article
German Research Consortium Unveils Open-Source AI Model Soofi
The German Association for Artificial Intelligence has officially launched Soofi S30B-A3B, a new open-source large language model. This milestone advances Europe’s sovereign AI infrastructure and revitalizes the open-source sector with a focus on hig
Toyota’s $6.4bn robotics estimate puts physical AI in focus
Toyota Motor projects that rolling out automation across its manufacturing facilities, subsidiaries, and key suppliers may require approximately 400,000 robots and an annual expenditure of roughly 1 trillion yen ($6.4 billion) starting in 2028.Accord
Meta Open-Sources 30B-Parameter Muse Glimmer for Local AI Agent on Consumer Hardware
On August 10, Meta launched Muse Glimmer, an open-source model designed to empower local AI agents on consumer hardware, marking a key step in its "Personal Superintelligence" strategy. With 30 billion parameters, this model is the open-source counte
Related Special Topic Recommendations
Comments (0)
0/500
Engineer Tibo, who oversees Codex and ChatGPT at OpenAI, shared news on X indicating that the 1 million-token context window feature of the GPT-5.6 Sol model within Codex is now accessible beyond API key holders and is also available to users of ChatGPT accounts. This change eliminates the previous limitation that forced subscribed users to cut off long-term project histories when dealing with large-scale tasks.

Subscribed users can manually activate the million-token context setting
Tibo explained that GPT-5.6 Sol in Codex now supports a context length of 1 million tokens, a capability previously reserved exclusively for API key holders. He noted that the feature has already been activated through ChatGPT accounts and that users with Plus, Pro, or other subscription plans can manually configure it to utilize the full 1.05 million-token capacity when working with the Codex programming agent.
A larger context window enables Codex to retain more code snippets, tool outputs, and conversation history before automatic compression kicks in, which is particularly beneficial for professional use cases such as large-scale codebase restructuring, intricate debugging, and extended sessions. For developers managing production-level repositories, this improvement directly addresses the previous issue of frequent context exhaustion.
The default setting was carefully chosen; enabling it requires careful consideration
Tibo also stressed that the current default context length represents the outcome of balancing performance needs with cost considerations rather than being set arbitrarily. He advised users to make an informed decision based on their own requirements. Several developers added that, thanks to Codex’s server-side compression technology, the default configuration is sufficient for most routine tasks.
Nevertheless, some users have reported that activating the million-token context setting leads to faster consumption of their subscription quotas, with a single intensive session potentially using up a significant portion of their weekly allowance. Experts recommend enabling this feature only when necessary rather than as the default setting for daily use, to prevent hitting usage limits prematurely on critical projects.
German Research Consortium Unveils Open-Source AI Model Soofi
The German Association for Artificial Intelligence has officially launched Soofi S30B-A3B, a new open-source large language model. This milestone advances Europe’s sovereign AI infrastructure and revitalizes the open-source sector with a focus on hig
Toyota’s $6.4bn robotics estimate puts physical AI in focus
Toyota Motor projects that rolling out automation across its manufacturing facilities, subsidiaries, and key suppliers may require approximately 400,000 robots and an annual expenditure of roughly 1 trillion yen ($6.4 billion) starting in 2028.Accord
Meta Open-Sources 30B-Parameter Muse Glimmer for Local AI Agent on Consumer Hardware
On August 10, Meta launched Muse Glimmer, an open-source model designed to empower local AI agents on consumer hardware, marking a key step in its "Personal Superintelligence" strategy. With 30 billion parameters, this model is the open-source counte











