K2 Think V2

MODEL RECORD: MBZAI-K2THINKV2

Organization
MBZUAI
Family
K2THINKV2
Model number
Unknown
Tier
Unknown
Status
ACTIVE
Release date
2025-12-15

Table may be scrolled horizontally on narrow displays.

Record No. Model Benchmark Metric Score Eval Date Source Model ID
BR-028K2THINKV2-001 K2 Think V2 Artificial Analysis Intelligence Index Overall 17.3 Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-002 K2 Think V2 GPQA Diamond Accuracy 71.3% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-003 K2 Think V2 Humanity's Last Exam No tools Overall 9.5% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-004 K2 Think V2 SciCode Overall 33% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-005 K2 Think V2 tau2-bench Overall 25.4% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-006 K2 Think V2 tau3-Banking Overall 5.4% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-007 K2 Think V2 Terminal-Bench Hard Overall 6.8% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-008 K2 Think V2 Terminal-Bench 2.1 Overall 15% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-009 K2 Think V2 CritPt Overall 0% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-010 K2 Think V2 AA-Omniscience Overall -33.9 Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-011 K2 Think V2 AA-Omniscience Accuracy 15.7% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-012 K2 Think V2 IFBench Overall 62.8% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-013 K2 Think V2 GDPval-AA Average Normalized Score 0% Unknown Primary source MBZAI-K2THINKV2
BR-028K2THINKV2-014 K2 Think V2 AA-LCR Overall 52.7% Unknown Primary source MBZAI-K2THINKV2