Inkling-Small — Real Scores Across 12 Benchmarks (2026)

Thinking Machines Lab

Inkling-Small

12 benchmarks

AIME

#22 of 117

95.1%

GPQA Diamond

#35 of 118

88.3%

IFBench

#1 of 3

83.4%

SWE-bench Verified

#23 of 86

77.4%

CharXiv

#20 of 23

76.7%

MCP-Atlas

#12 of 17

74.9%

MMMU Pro

#26 of 36

73.1%

SWE-bench Pro

#31 of 43

53.2%

Terminal-Bench 2.1

#27 of 27

52.7%

Humanity's Last Exam

#26 of 72

46.6%

Humanity's Last Exam

#26 of 72

29.6%

SimpleQA Verified

#5 of 5

20.9%