List your company

Scientific Question Answering on GPQA Diamond (test)

49Pass@1Transformers
37.14440.22243.346.378Jan 26, 2026

Evaluation results

MethodLinks
TransformersBackbone=R1-Qwen2026.0149349.4
H2OBackbone=R1-Qwen2026.0148.1593.6
DYNTSBackbone=R1-Qwen2026.0148.1625.7
R-KVBackbone=R1-Qwen2026.0147.2615.1
SnapKVBackbone=R1-Qwen2026.0146.5622
TransformersBackbone=R1-Llama2026.0146.4207.9
DYNTSBackbone=R1-Llama2026.0146.3397.6
WindowBackbone=R1-Qwen2026.0145.9634.1
StreamingLLMBackbone=R1-Qwen2026.0145.9631.1
SepLLMBackbone=R1-Qwen2026.0145.6639.5
R-KVBackbone=R1-Llama2026.0144.5394.9
SnapKVBackbone=R1-Llama2026.0141.9394.1
H2OBackbone=R1-Llama2026.0141.6381.5
SepLLMBackbone=R1-Llama2026.0139.7406.6
StreamingLLMBackbone=R1-Llama2026.0137.8407.4
WindowBackbone=R1-Llama2026.0137.6408.8