AI Week in Review 26.10.10 - by Patrick McGuinness

The review states that the tier is available across OpenAI's API, Codex and ChatGPT Work platforms. According to the review, the tier was engineered for latency-critical agent workflows. The publication is titled 'AI Week in Review 26.10.10.'

The review reports that the Ultrafast tier delivers an 8 to 14 times throughput boost while maintaining near-Astra frontier reasoning capabilities. It also states that OpenAI halved top-tier inference API pricing. The review additionally reports that OpenAI made automated code review tooling free for developers. These three items — the new service tier, the price reduction and the free tooling — are presented together as the week's principal OpenAI news. No further breakdown of the throughput range is provided.

The review presents the changes as product and pricing updates relating to the GPT-6.1 Sol model family. It frames the throughput boost and the pricing change as updates to OpenAI's developer-facing services across the API, Codex and ChatGPT Work. The excerpt does not include further detail on availability dates, regional limits, or the terms attached to the free automated code review tooling. The review characterises the throughput gain as intended for latency-critical agent workflows, set alongside the halving of top-tier inference API pricing and the removal of charges for automated code review.