Science
Google topped AI model rankings as OpenAI training remained paused
Google led the field while OpenAI's Astra and GPT-6.1+ releases were delayed.
On Saturday, Google held the top position in AI model performance rankings for the month. The assessment placed Google ahead of Anthropic, OpenAI and Meta. OpenAI had Not resumed training on its models by mid-October. Consequently, the expected Astra 6.1+ and GPT-6.1+ releases did Not appear in October.
The ranking evaluated models released up to the end of October. Google's system scored highest across benchmark tasks. Anthropic, OpenAI and Meta followed in the list.
OpenAI's training operations remained halted through October 20. Earlier checkpoints on October 14, 16 and 18 also showed no restart. The pause extended the timeline for any new model development.
The Astra model, version 6.1 or higher, was Not launched by October 31. No public announcement or deployment of that version occurred during the month.
Likewise, the GPT-6.1+ Astra variant was absent from October releases. No version matching that designation was made available in the month. This reinforced the gap between Google's active offerings and OpenAI's delayed pipeline.
Analysis
By outperforming rivals on standard benchmarks, the company gains credibility with enterprise customers and may attract further investment in its research pipeline.
OpenAI's continued training pause delays the rollout of its next-generation models, which could affect developers who rely on its APIs. The lack of new releases may also give competitors additional time to close performance gaps, altering the competitive dynamics for the remainder of the year.
Stakeholders must now consider a longer horizon for new capabilities from OpenAI, while Google may consolidate its lead unless it encounters its own development hurdles.
What this was built from
Each claim above traces to a market below, at the price it was trading when this edition went to press.