Claude Haiku 5.5 (Max) vs GPT-6 Astra (Medium)

What the captured data shows

Claude Haiku 5.5 Max has lower listed token prices than GPT-6 Astra Medium. Astra Medium has the higher intelligence index and a much shorter reported first-token delay, while Haiku has higher reported output throughput. The Max and Medium configurations are important parts of the evidence rather than interchangeable labels.

Which model fits your workflow

Check the time to a completed, acceptable answer on a representative task. A throughput advantage may be less useful if reasoning delays the first response. For unattended work, compare the total completion time and the number of attempts needed to satisfy the requirement.

Decision takeaway

The captured data supports a lower listed Haiku cost, but the different reasoning configurations and timing measurements need to stay visible.

These figures come from Franklin’s Artificial Analysis-sourced model records captured on 2026-10-11. Read the individual model entries for the data snapshot: Claude Haiku 5.5 (Max) and GPT-6 Astra (Medium). Missing measurements remain unavailable; this page does not report a new hands-on test.