OpenAI Raises GPT-6.1 Sol Default Context to 500,000 Tokens
The default context length for GPT-6.1 Sol in Codex and ChatGPT Work rose from 272,000 tokens, an increase of about 80%.
OpenAI has raised the default context length of GPT-6.1 Sol from 272,000 tokens to 500,000 tokens when the model is used through Codex and ChatGPT Work, an increase of about 80%. Tibo Sottiaux, an OpenAI product lead, revealed the change on October 12, according to aitimes.com.
The new default is intended to reduce how often context compaction occurs during long tasks, so users spend less time working around the limit. Users can still adjust the context length manually, raising it above 500,000 tokens or lowering it to save resources.
GPT-6.1 Sol supports a maximum context length of 1 million tokens, leaving room above the new default. Sottiaux said the default was set after sufficient testing and that OpenAI will expand the default context length further in the future.
Part of a wider Codex push
OpenAI is in the middle of a 28-day Codex performance improvement event, and the context change is the day 7 update. The company has not said what the remaining days of the event will cover.
Quick answers
What is the new default context length for GPT-6.1 Sol?
It is now 500,000 tokens when the model is used through Codex and ChatGPT Work, up from 272,000 tokens.
What is the maximum context length of GPT-6.1 Sol?
The model supports a maximum context length of 1 million tokens, and users can adjust the setting manually.
Who announced the change?
Tibo Sottiaux, an OpenAI product lead, revealed the change on October 12.