Fan-made chart claims MAI Code 2 matches 96% of Claude Opus 5.5 coding perf at 5% of the cost
marlene_zw · x · 2026-11-20
X user marlenezw predicts "MAI models will lead the absolute frontier next year," attaching a fan-made coding benchmark chart. It claims MAI Code 2 averages 63.2% pass rate across DeepSWE 1.1, Terminal-Bench 2.1, SWE Atlas and FrontierCode v1.1 — about 98% of GPT-6.1 Sol (64.1%) and 96% of Claude Opus 5.5 (65.3%) — at $0.309 average cost per task vs $0.761 and $6.306 respectively. The chart references unreleased model versions and is unofficial and unverified; treat it as community optimism around Microsoft's MAI series rather than a real benchmark.
Related event: Microsoft launches MAI Code 2 models, brings them to GitHub Copilot(3 posts)→
More from Models
- Claude models 82% off on RouterHub: Opus 5.5 at $0.72 in, Fable 5.2 at $1.43 in — zebird0 · 2026-11-20
- Team claims Codex cracked 35 open Erdős problems in 2 months, some Lean-verified — IgorCarron · 2026-11-20
- Google's new AI reportedly cheaper and faster than Claude Opus 5.5 — HotMinuteUK · 2026-11-20
- GPT-6.2 Astra tops Blueprint-Bench 2, claims best physical 3D understanding among AI models — gdb · 2026-11-20
- MiniMax H3 video model reportedly blocks explicit prompts despite reference images — silleeni · 2026-11-20
- Google ends free Gemini Pro student subscriptions worldwide as model matures — Angaisb_ · 2026-11-20