https://grindbench.ca/New models, method changes and corrections from GrindBench.https://grindbench.ca/findings/we-marked-our-own-answer-key/2026-11-03-We audited our Banking answer key and fixed four errorsTue, 03 Nov 2026 12:00:00 GMTWhere three or more models agreed on an answer our key marked wrong, we checked the key by hand. Of 17 such answers, the key was right 13 times and wrong 4 times. We fixed the four, regraded every model from its saved answers, and published the audit. Scores moved by at most 4 points.https://grindbench.ca/rankings/2026-11-02-Eight more models on BankingMon, 02 Nov 2026 12:00:00 GMTGemma 4 26B and 31B, Mistral Small 4, Nemotron 3 Super, Ultra and 3.5 Lightning, Qwen3.6 35B and Qwen3.8 Flash. Eighteen models now have results on both benchmarks.https://grindbench.ca/about/#limits2026-11-02-Banking's standard run is now 60 tasksMon, 02 Nov 2026 12:00:00 GMTTen tasks from each of six kinds of banking work, chosen from the original 150. On the ten models that ran all 150, the shorter run changed no score by more than 3.3 points and costs 40% as much. Scores for single kinds of work now rest on 10 tasks each and have wide ranges.https://grindbench.ca/rankings/2026-11-01-GrindBench-BankingSun, 01 Nov 2026 12:00:00 GMTA second benchmark: the everyday analyst work of Canadian banks and credit unions, from anti-money-laundering reports to capital ratios. Its ranking of models agrees closely with the public-sector one.https://grindbench.ca/models/kimi-k3-1-pro/2026-10-29-Kimi K3.1 Pro on PubSecThu, 29 Oct 2026 12:00:00 GMTTwenty-eight models now have public-sector results.