Locus ranks 4th on average across six Kaggle competitions
mariofilhoml · x · 2026-08-04
The post says Locus’s capabilities extend beyond post-training, and highlights results verified with the PostTrainBench authors.
According to the quote, Locus was run out of the box on all prize-money Kaggle competitions with public leaderboards that were active at the time, with no competition-specific setup or instructions. It reportedly achieved the 4th highest average performance across six competitions and outperformed an average of 89.5% of human competitors. The quote also says Locus is being used in domains such as quantum chemistry and speech recognition.
The author adds a caveat that the “4th average rank” is a somewhat creative marketing-style calculation, while still calling the results impressive.
Related event: Automated System Locus Sets New Post-Training Records(4 posts)→
More from Models
- Epoch AI’s MirrorCode benchmark sees Claude Fable solve C preprocessor and Pkl tasks — Jsevillamol · 2026-08-04
- X post jokes that Anthropic could hit $100B in revenue by year end — Jsevillamol · 2026-08-04
- DeepSeek chatter says the company is moving beyond a V4-Ultra-style model — teortaxesTex · 2026-08-04
- Windows reset restores RTX 3070 throughput for local Qwen3.6-35B inference — campaigner_ · 2026-08-04
- ChatGPT’s anti-sycophancy training may be causing “performative nuance” — Chance-Physics-7216 · 2026-08-04
- Qwen 3.8 Max scores well on analysis tasks, Reddit says it is a cheap Opus 4.7 replacement — Ill_Distribution8517 · 2026-08-04