Kimi k3 ranks second on the agentic benchmark but costs more than opus 4.8 to run and averages…
kimi k3 ranks second on the agentic benchmark but costs more than opus 4.8 to run and averages nearly an hour per task. frontier quality is now measured in dollars and hours per task, not accuracy.
the leaderboard that matters is your invoice.
Context
Artificial Analysis's article of 21 July 2026 reports AA-Briefcase, its proprietary agentic knowledge-work benchmark: Kimi K3 scores Elo 1543, second to Claude Fable 5 at 1574, ahead of GPT-5.6 Sol (max) at 1501, Claude Sonnet 5 (max) at 1388 and Claude Opus 4.8 (max) at 1347. Kimi K3 costs 10.57 dollars per task, which AA says is more than Opus 4.8, and averages 56.4 minutes, 83 turns and 120k output tokens per task.
Second is as of AA's 21 July article, two months before the note, and the current leaderboard was not checked. The figures are one operator's harness and config and are not merged with other benchmarks. The exact Opus 4.8 per-task cost was not extracted.
Related work
- AA-Briefcase evaluation (Artificial Analysis) ↗Linked from the article; not read.
Watch next
- The live AA-Briefcase page for later rankings.
Sources
- Kimi K3 agentic knowledge benchmark (Artificial Analysis, 21 Jul 2026)artificialanalysis.ai
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 18 September 2026 at 04:31 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 →