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

What the captured data shows

Claude Haiku 5.5 Low has lower listed token prices and higher reported output throughput than GPT-6 Astra Medium. Astra Medium has the higher intelligence index and a slightly shorter reported first-token time in this snapshot. These signals answer different questions and should not be collapsed into one speed score.

Which model fits your workflow

For a short response, first-token time can matter more than output throughput. For a longer answer, record both and inspect whether the answer meets the requirement. Preserve each reasoning setting when reproducing the comparison, and account for retry costs when evaluating the cheaper entry.

Decision takeaway

The recorded data supports separate comparisons of price, first-token delay and throughput, rather than a universal winner.

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 (Low) and GPT-6 Astra (Medium). Missing measurements remain unavailable; this page does not report a new hands-on test.