← Leaderboard
unsloth/qwen3.6-35b-a3b

Qwen3.6 35B Q4 (llama.cpp)

llama.cpp MoE · 3B active / 35B total AgentsGeneralReasoning
Throughput
48.6t/s@ 32k
Engine
llama.cpp
Parameters
3B active / 35B total
Released
2026-04-16
Benchmarked
2026-06-27

Context ladder

Throughput at each benched context window (single measurement).

Context KV Throughput
@ 32k peak golden 48.6t/s

Golden profile

qwen36-q4-llama

Capabilities

reasoningggufllamacppgolden

Why we run it

Golden fleet target — auto-scaffolded from recipe qwen36-q4-llama.

Bench notes

golden ?/? @ 38.8 tok/s — fill~14745 — bench-agent-v2 — tool_ok=False

Benchmarked 2026-06-27
SparkBench · GB10 · single node