tyrtleli/thinking-benchmark-gpt5-5 download history

tyrtleli/thinking-benchmark-gpt5-5 is a question answering dataset on the Hugging Face Hub. In the last 30 days it was downloaded 15 times (2 in the last 7 days), and 96 times in total. It ranks #450,867 among datasets by monthly downloads.

Thinking Benchmark — GPT-5.5 k=32 subset Note on saturation: GPT-5.5 achieves 96.6% overall accuracy on the 85-problem calibration pool (78/85 problems solved 8/8 at k=8). Only one problem falls in the classic 25–75% edge band. This 6-problem set uses the widened 1–7/8 range (per the

Open tyrtleli/thinking-benchmark-gpt5-5 on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.