K-Bench: measuring model performance on real scientific agent requests

K-Bench 01, an evaluation built from first-turn requests sampled from live user traffic on K-Dense Web, ran 1,602 completed agent runs by nine frontier models in identical sandboxes, with three blinded language-model judges scoring each run against an eight-dimension rubric. No model clears the acceptance threshold under all three judges; gpt-5.6-sol has the highest pooled mean of 8.04, but its 95% interval [7.80, 8.23] spans the threshold, and two judges rank claude-opus-5 first. Across 39,934 scored judgments, 47.6% fall below the 8-point threshold, with scientific accuracy averaging 6.22 versus 7.33 for communication, and overclaiming the leading failure tag on 31.4% of assessments.

arXiv:2608.21601v1 Announce Type: new Abstract: Benchmarks for scientific artificial intelligence are mostly written to be scored: multiple-choice questions, curated agent tasks with reference solutions, or simulators with a known generative structure. Real scientific requests arrive differently. They are underspecified, they carry attachments, and lack ground truth. We report K-Bench 01, an evaluation built from first-turn requests sampled from live user traffic on K-Dense Web and run end to end by nine frontier models in identical sandboxes, yielding 1,602 completed agent runs. Three blinded language-model judges scored every run against an eight-dimension rubric. On a rubric whose 8-anchor instructs judges that a domain scientist would accept the work with minor edits, no model clears the line under all three judges. gpt-5.6-sol has the highest pooled mean, 8.04, but its 95% interval 7.80, 8.23 spans the threshold, and two of the three judges rank claude-opus-5 first instead. We therefore report the ordering of systems as the reproducible quantity, the absolute level as an attribute of the instrument, and the top of the table as unresolved. Across all 39,934 scored judgments -- the eight dimension scores plus a holistic overall for each assessment, excluding not-applicable cells -- 47.6% fall below the 8-point threshold. Difficulty is not uniform across the rubric: scientific accuracy averages 6.22 against 7.33 for communication, on identical denominators and in the same direction within every one of the nine models. The single leading failure tag is overclaiming, on 31.4% of assessments. We argue that the informative quantity for scientific agents is not a leaderboard position but the joint distribution of what was delivered, what was claimed, and what artifacts were produced.