FrontierSWE leaderboard: GPT-6 Astra tops at 65.5%, Gemini 4 Carbon scores 61.2% at 1/16th the cost
scaling01 · x · 2026-10-29
ProximalHQ's FrontierSWE leaderboard (34 tasks, 5 trials each, 20-hour budget) ranks GPT-6 Astra first at 65.5%, followed by Claude Opus 5.5 (62.3%), Claude Sonnet 5.5 (61.9%) and Gemini 4 Carbon (61.2%), with Claude Fable 5.1 (56.3%) and Gemini 4 Argon (55.0%) behind. Costs vary wildly: GPT-6 Astra averages $1,029.65 per run vs $61.93 for Gemini 4 Carbon, making the latter far cheaper per point.
More from Models
- Opus 5.5 reportedly notably stronger than Astra at writing custom kernels — DimitrisPapail · 2026-10-29
- ChatGPT unifies Chat and Work — users wonder how usage limits will change — Angaisb_ · 2026-10-29
- Replit's ViBench 2.0: top models pass only 49-65% of long-horizon builds — amasad · 2026-10-29
- Replit's ViBench 2.0: best models pass only 49–65% of builds, GPT-6.1 Sol leads — amasad · 2026-10-29
- OpenAI now includes monthly API credits with ChatGPT Pro and Business subscriptions — Angaisb_ · 2026-10-29
- Ex-Google Researcher's Elorian Launches Glimpse: Frontier Models Fail 50% on Visual Counting Benchmark — Latent Space · 2026-10-29