List your company

Scientific Reasoning on SciKnowEval Hardest

21Pass@1STV verifier
0.725.98511.2516.515May 28, 2026

Evaluation results

MethodLinks
STV verifiergenerator=Qwen3-8B2026.0521
Qwen3-32Bsize_multiplier=4x2026.058.9
Qwen3-235B-A22Bsize_multiplier=30x2026.058
No verifier traininggenerator=Qwen3-8B2026.052.1
No verificationgenerator=Qwen3-8B2026.051.5