VulcanBench(@VulcanBench):I got a LOT of positive feedback about this concept. So I built the eval suite, Routine v1, and Opus 5.5 will be the first model tested on it. Sample below of what’s coming once I have a handful of models tested on it. First of its kind, not trying to stump the frontier or solve any crazy hard unsolvable proofs. Trying to show team what model and effort level they actually need to move from tokenmaxxing to tokenminning 🖖

VulcanBench

@VulcanBench

Benchmarking models across effort levels and harnesses on eval suites with real coding tasks. Because you don't need High or Max effort as much as you think.

加入 March 2020

71 正在关注 2.7K 粉丝

VulcanBench@VulcanBench

2026.09.25 00:25

I got a LOT of positive feedback about this concept. So I built the eval suite, Routine v1, and Opus 5.5 will be the first model tested on it. Sample below of what’s coming once I have a handful of models tested on it. First of its kind, not trying to stump the frontier or solve any crazy hard unsolvable proofs. Trying to show team what model and effort level they actually need to move from tokenmaxxing to tokenminning 🖖

显示更多