OpenAI Unveils Nearly 50% Boost in GPT-6 Speed, Yet Third-Party Tests Show Its Subscription Quotas Lag Far Behind Claude's
2 hour ago / Read about 0 minute
Author:小编   

Tibo has either released or reset its commitment fulfillment for 28 days straight. On the inaugural day, OpenAI heralded a nearly 50% surge in the operational velocity of GPT-6 Astra and GPT-6.1 Sol, with enhancements spanning across all platforms and integrated products from associated partners. Users are set to witness these alterations within a two-hour window, sans any necessary intervention. Concurrently, third-party entity SemiAnalysis unveiled a pragmatic test report, highlighting that the API equivalent value of Anthropic's subscriptions is roughly quintuple that of OpenAI's, with a notably wide chasm at the mid-tier model level, which is commonly utilized on a daily basis. The report also elucidates that following OpenAI's adjustment of its subscription schemes, the quota for the $200 tier has been slashed by half, whereas the newly introduced $500 tier for premium models merely offers a 21% increment in quota compared to the erstwhile $200 tier, thereby eradicating the prior rationale of enhanced value for superior tiers. Both OpenAI and Anthropic are curtailing subscription subsidies, albeit through disparate methodologies: OpenAI has directly diminished quotas across the board for all models, while Anthropic has incrementally diminished the subscription value of its top-tier models. Presently, the speed enhancements have garnered some acclaim, yet the quota reduction remains an undeniable fact. Upon the expiration of old quotas on October 29, heavy users will promptly sense the repercussions of the diminished quotas. Over the ensuing 27 days, OpenAI will be tasked with mitigating user discontent regarding the quotas.