GPT-5.6 Sets New Record on ALE
shi_weiyan · x · 2026-07-11
OpenAI stated that GPT-5.6 Sol scored 53.6 on Agents' Last Exam (ALE), outperforming Claude Fable 5 (adaptive) by 13.1 points. Under medium reasoning settings, it beat Fable 5 by 11.4 points at roughly 1/4 the estimated cost; GPT-5.6 Terra and Luna also surpassed Fable 5 at about 1/16 the cost.
The post also provided background on ALE: a public benchmark for frontier AI agents featuring 1,500+ expert-provided, long-horizon tasks. It covers 55 non-embodied professions, with input from 300+ experts across 100+ institutions. The benchmark emphasizes real-world professional work, reproducibility, and automated outcome-based grading.
Related event: GPT-5.6 Sets New Record on ALE Benchmark(2 posts)→
More from Models
- BullshitBench update: GPT-6-Astra beats all prior OpenAI models but still trails Anthropic — scaling01 · 2026-09-11
- Astra Scores 83% on GauntletBench, First Computer-Use Agent to Beat Human Baseline — ducha_aiki · 2026-09-11
- Kimi K2.8 Preview rolls out: near-K3 coding performance, 1M context for all tiers — teortaxesTex · 2026-09-11
- Looking for a classifier of software engineering task shapes to pick models per task — StewartalsopIII · 2026-09-11
- DeepSeek V4 Pro API to continue after Sept 2026, billing unchanged — teortaxesTex · 2026-09-11
- DeepSeek V4.1 Flash Hits 98% of GPT-6 Astra's Score at 1.4% of the Cost in Third-Party Benchmark — ayushtweetshere · 2026-09-11