all the models — AI benchmark observatory
← Benchmarks

PinchBench

PinchBench evaluates coding agents on real-world agentic coding tasks, measuring both best-case and average performance across complex software engineering scenarios.

id pinchbench · max 1 · 6 models reported

#ModelScore
1Nemotron 3 Ultra (550B A55B)
NVIDIA · open
0.90
2Nemotron 3.5 Lightning (30B A3B)
NVIDIA · open
0.85
3MiMo-V2-Omni
Xiaomi
0.81
4MiMo-V2-Pro
Xiaomi
0.81
5GLM-5V-Turbo
Zhipu AI
0.81