According to Decrypt, leaks indicate that Anthropic may be preparing a new version labeled Mythos 5.5, while OpenAI appears to be testing a model named GPT 6.2 Astra. Users have reported observing requests being routed through an unidentified stronger model during interactions with Claude. The label “Mythos 5.5” was found in Anthropic’s website code, according to the source. No official announcements or timelines were mentioned. OpenAI’s testing of GPT 6.2 Astra remains preliminary, per the report.
