Skip to content
brianletort.aiD. Brian Letort, Ph.D.

AI Market · Forecast desk · miss · software

60% · miss

Gemini 3.5 Pro reaches public GA by June 30, 2026, but does not exceed Claude Opus 4.8 on SWE-Bench Pro in its first independent Artificial Analysis run.

Back to the bookIssue 2026-W23Evidence ↗

Outcome

misscalled at 60%resolved July 18, 2026brier contribution 0.360timing-slip

GA premise failed: Gemini 3.5 Pro had not shipped by June 30 — the public Gemini API still listed only gemini-3.5-flash and gemini-3.1-pro as of mid-July — so the SWE-Bench Pro comparison leg was never tested.

Evidence: https://ai.google.dev/gemini-api/docs/models