The agent exam just graded everyone below 35 percent. argo-bench, out oct 2, scored claude opus 5.5…
the agent exam just graded everyone below 35 percent. argo-bench, out oct 2, scored claude opus 5.5 as the top agent and it still cleared only 34.8 percent of 210 real tasks, averaging 59.5 points.
the passing bar now sits above every model.
Context
Verified dates: the Argo-Bench paper, 'Evaluating Data Agents on Enterprise-Scale Workflows' (arXiv 2610.02122), was published Oct 1, 2026 by TextQL researchers, and AI.info covered it on Oct 2. It is a 210-task benchmark built on a simulated New York City food delivery platform, with ground truth hidden in an Oracle E-Business Suite warehouse of 235 tables and 7.5 billion rows.
TextQL says the strongest model, Claude Opus 5.5, fully solves 34.8 percent of tasks, scoring 95 or more, and averages 59.5 points; eight of the thirteen models average under 30.
The 34.8 percent, the 59.5 average, the 210 tasks and Claude Opus 5.5 at the top match TextQL's page and the paper coverage. The post says the benchmark was out Oct 2; the paper is dated Oct 1 and Oct 2 is the AI.info news date. The sources differ on model count: TextQL says thirteen models, AI.info says 14 frontier and open-weight models. 'Solved' means a score of 95 or more, so the 'passing bar' is the paper's threshold. 'The passing bar now sits above every model' is the author's opinion.
Related work
- The Aggregate: Argo-Bench leaderboard ↗Tracks 43 models with a saturation forecast.
- eWeek: top AI model fully solves just 35% of enterprise tasks in Argo-Bench test ↗Oct 4, 2026.
- AI.info: Argo-Bench paper summary ↗Oct 1, 2026.
Watch next
- Check the leaderboard for the v1.1 run and any new models added after Oct 2.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 11 October 2026 at 16:41 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →