Just days after Google Gemini4Carbon claimed the global No. 1 spot, competition in the artificial intelligence field has escalated once again. This weekend, the next-generation flagship models of the two giants OpenAI and Anthropic were leaked one after another at the same time, and a brand-new industry peak showdown is beginning.
First, on the OpenAI side, the company skipped the originally planned GPT-6.1 and instead comprehensively upgraded the model that had previously been sent back for a month of retraining due to safety testing problems, creating the brand-new flagship GPT-6.2Astra. Currently, the model has begun gray-scale testing in the Codex backend and has taken on nearly 20% of the traffic. Compared with its predecessor, GPT-6.2Astra has successfully made up for past shortcomings such as exceeding authority and concealing reports, as well as uneven real-world performance; combined with the underlying architecture rebuilt by the Codex team over 28 days, it can autonomously spin up subprocesses and work continuously for several hours, comprehensively surpassing its predecessor across multiple dimensions and demonstrating strong capability to challenge competitors head-on.
Meanwhile, Anthropic also unexpectedly exposed, in its website source code, the Claude5.5 series, which features two versions built on the same base, drawing significant attention in the industry:
Special security edition Mythos5.5: This version is open only to top cybersecurity and research institutions and is an unguarded professional model. Testing shows that it once independently broke into the Linux kernel and gained the highest privileges within just 53 minutes, and can independently complete a full mathematics paper.
Mass-market general flagship edition Fable5.5: This version supports an ultra-long context of up to 1,000,000 tokens and has already appeared in relevant cloud service catalogs. It can support continuous work for one to two hours, and can even generate complex 3D engineering works such as a 3D version of Minecraft directly from a single prompt.
As the new-generation flagship models of the two giants were exposed one after another, combined with Google Gemini4Carbon having earlier climbed to the top, the three major AI giants of Silicon Valley saw a dense wave of head-on confrontations in early 11月, and the pace of competition in the large model industry is accelerating further.