AI2 chart: GPT-5.5 over-helps students 83% of the time, Claude Sonnet 5.5 just 33%
allen_ai · x · 2026-10-13
Allen AI's TutorMoments chart shows how often AI tutors over-help (lower is better): Claude Sonnet 5.5 at 33%, GPT-6 Astra 35%, GPT-6 Sol (no reasoning) 43%, Claude Opus 5.5 49%, Claude Opus 4.8 54%, Claude Sonnet 4.6 56%, versus 82%/83% for GPT-5.5 (no reasoning) and GPT-5.5. GPT-6 models over-help far less than earlier OpenAI models and score higher overall when explicitly instructed to let students work through what they can handle.
Related event: Allen AI's TutorMoments Benchmark Finds LLM Tutors Overdo It for Students(5 posts)→
More from Models
- Announcing AA-3D: Artificial Analysis benchmarks AI models building interactive 3D scenes — ArtificialAnlys · 2026-10-13
- Same models top almost every AA-3D domain, Artificial Analysis finds — ArtificialAnlys · 2026-10-13
- Four models form the AA-3D cost frontier, from $0.20 to $15.18 per scene — ArtificialAnlys · 2026-10-13
- Scale's Alexandr Wang says muse-wonder bug fixes are already showing clear improvement — alexandr_wang · 2026-10-13
- Muse's $15 plan tested: roughly 1B tokens a month, ~$280 at API rates — Aldarund · 2026-10-13
- Fine-tuning speech models for Algerian Darija: a team's four open questions — Fair_Bed_914 · 2026-10-13