Surge AI launches DAYJOB benchmark: end-to-end professional judgment with 14-17 hour work horizons
echen · x · 2026-09-24
Surge AI introduced DAYJOB, a benchmark suite measuring whether agents can perform economically valuable professional work end-to-end—figuring out what needs doing, making informed decisions, and reasoning over days—beyond single clear tasks.
Compared with GDPval, DAYJOB reshapes the tasks:
- Prompts shrink from 337 to 49–81 words; environments are messier
- Working environments grow from 1.2 to 20–26 files
- Estimated human work grows from 4 to 14–17 hours
Early results show frontier agents still have a long way to go.
Related event: Surge AI Releases DAYJOB Benchmark; Top Agents Pass ~25%(3 posts)→
More from Research
- MIT's Tegmark coins 'formal-washing': Signal protocol proof is far from complete — tegmark · 2026-11-12
- Signal formal verification hits 80%, Tegmark says Wired overstated completion — tegmark · 2026-11-12
- Category-theoretic framework asks: when does a physical system actually compute? — neurovium · 2026-11-12
- Radical Numerics brings genome language model Omnii-1 to the edge for biodefense — exnx · 2026-11-12
- Against 'verifiable vs. fuzzy': the real axis is how expensive tasks are to grade — dfrsrchtwts · 2026-11-12
- Toby Ord flags 'all-things-considered' AGI forecast table is actually conditional on no pacing — tobyordoxford · 2026-11-12