Livecodebench Websites
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Lan
LiveCodeBench: Holistic and Contamination-Free Evaluation of Large Language Models for Code
Not Applicable
$
8.94
K2 Think — Open‑Source Reasoning System (32B) · Blazing 2,000 tok/s ·
K2 Think is a 32B open‑weights reasoning system from MBZUAI & G42. Apache‑2.0 license. Built on Qwen2.5‑32B, tuned with long CoT SFT + RL with verifiable rewards, agentic planning, and test‑time scaling. Up to ~2,000 tok/s on Cerebras WSE. Benchmarks: AIME’24 90.83, AIME’25 81.24, HMMT’25 73.75, OMNI‑MATH‑HARD 60.73, G
Not Applicable
$
8.94
