Epoch AI adds latency data for GPT-6 Astra's long-context pricing

Epoch AI @EpochAIResearchOriginally posted 2026.09.11

Open original on X

Earlier this week, we showed that GPT-5.6 and Claude 5 exhibit very different long-context TTFT scaling, which corresponds to their different pricing structures at long context lengths. OpenAI recently released GPT-6 Astra, which shows the same pattern of increased API pricing beyond 272k input tokens. We collected additional latency measurements, which show a similar curvature to that of GPT-5.6 models.

This post quotes another post: x.com/1529761561170124800/status/209743297517256…