GPT-6.2 Astra hits ~78% on Agents' Last Exam at fraction of rivals' cost
adonis_singh · x · 2026-11-18
A viral "Agents' Last Exam v1.1" cost-performance chart shows OpenAI's new GPT-6.2 Astra reaching 78% at roughly $2-3 of cost — matching Gemini 4 Carbon's score while undercutting Claude Fable 5.2 ($60) and Claude Opus 5.5 ($15) by a wide margin. GPT-6.2 Sol also beats the previous GPT-6.1 Sol at near-zero cost. The chart suggests GPT-6.2 shifts the cost-quality Pareto frontier dramatically leftward.
Related event: GPT-6.2 Astra Tops Agents' Last Exam at Fraction of Cost(4 posts)→
More from Models
- GPT-6.2 Astra models RC buggy part by part, generating animated assembly guide and parts catalog — Dr_Singularity · 2026-11-18
- Microsoft launches MAI Cyber 2 Flash, beats GPT-6 Astra on CyberGymE2E at 2% of the cost — mustafasuleyman · 2026-11-18
- Microsoft-Decision-1.1 hits GA with 2x speed at lower cost — marlene_zw · 2026-11-18
- Microsoft announces MAI Code 2 for GitHub Copilot — testingcatalog · 2026-11-18
- Professor: using AI detectors as a static prior is just confirmation bias — ipeirotis · 2026-11-18
- Wikipedia-game benchmark: GPT-6.2-Sol wins computer-use speed test at 39 seconds — dotey · 2026-11-18