@kaggle
The leading platform for crowdsourced benchmarks, competitions, and hackathons where the community shapes real-world progress in AI.
Kaggle previews Evaluated’s first episode, with Gina Lee hosting SWE-bench co-author Ofir Press on benchmark-building.
ScarfBench checks whether coding agents preserve app behavior across 204 migrations among Jakarta EE, Quarkus and Spring.
In an October 28 post, echen said the planned San Francisco talk would address ambiguous prompts and professional judgment.
The FrontierCS 2 announcement says agents must outdo each paper’s original method without access to the paper or web.
Kaggle says base models tend toward the mean rather than reflecting how preferences vary between cohorts.
Kaggle says entrants will classify passages from 18 models across six organizations, with a $50,000 prize pool.
Kaggle asks people to pick one competition or hackathon they'll never forget.
The Nobel Prize account cites his IceCube work and the discovery of high-energy neutrinos of astrophysical origin.
Kaggle asks entrants to find actionable links between 10 Hz Combine sensor data and future regular-season NFL performance.
Kaggle is asking people to share their best hackathon advice.
Kaggle is encouraging the community to export AutoScientist open weights to its platform.
Kaggle says the contest, hosted with Google Gemma, challenges entrants to post-train an open model into a coding agent for everyday hardware. Entries are due November 25, 2026.
HyperFrames says it built the benchmark with Google DeepMind and Kaggle, drawing a distinction between coding for SWE-bench and generating video that “feels alive.”
Kaggle asks whether computing scale has won out over clever engineering on a small budget.
Kaggle says the Enveda CASMI 2026 competition asks entrants to build machine learning models that predict 2-D chemical structures from mass spectrometry data. The entry deadline is December 7, 2026.
New benchmark tests schema-guided extraction on complex enterprise documents and scans.
Partnership with GertLabs creates a two-sided game testing AI security in customer support.
