//opengauntlet
← back to the leaderboard

Qwen 3.6 27B Architect Polaris Fable NVFP4 conversational AI benchmark

qwen36-27b-architect-polaris-fable-nvfp4

Qwen 3.6 27B Architect Polaris Fable NVFP4 benchmark: humanlikeness scores, pairwise rankings, transcripts, latency, and measured local speed.

Leaderboard rank
#16 of 41
Humanlikeness
67.1 / 100
Pairwise rating
1619
Measured speed
9.8 words/sec

Evidence: identical scenario pack and judge protocol; measured and judged results are shown separately below. Read the methodology.

Loading model…