← Leaderboard

thinkingcap-qwen3.6-27b

protolabsai/thinkingcap-qwen3.6-27b protolabsai/thinkingcap-qwen3.6-27b — gated or private
Q5_K_M 27B llama.cpp General

Recipe

Profile
protolabsai-thinkingcap-qwen3-6-27b-llama
Engine
llama.cpp
Context
32k · q8_0 KV
Served as
thinkingcap-qwen3.6-27b

Run on your Spark

single node shell
spark inference up protolabsai-thinkingcap-qwen3-6-27b-llama

Bench notes

bench-v2 avg 6.8 decode tok/s (2 sessions, ~50k ctx fill, tool_ok=True)

Benchmarked 2026-07-11
SparkBench · GB10 · single node