Can an RTX 4090 run Qwen3.6-35B-A3B?
Yes: an RTX 4090 runs Qwen3.6 35B-A3B at 160.6 tok/s at Q4_K_M, measured on llama.cpp with a 112-token prompt in October 2026.
1 signed run on llm-speed. Single-request decode speed in tokens per second; the short-chat workload unless noted. This run was measured by llm-speed (@llmspeed). Runs measured by llm-speed are credited to @llmspeed.
Measured tok/s by quant and runtime
| Quant | Runtime | Build | Decode tok/s (best run) | Workload | Runs | Source |
|---|---|---|---|---|---|---|
| Q4_K_M | llama.cpp | 8931 (9725a313b) | 160.6 | chat-short | 1 | Measured by llm-speed@llmspeed |
Fastest quant that fits: Q4_K_M on llama.cpp, 160.6 tok/s (measured by llm-speed; it fits because it ran on this hardware). Quants without a run are not listed.
Every measured row
| Run | Date (UTC) | Workload | Quant | Runtime / build | Prompt tokens | Output tokens | Decode tok/s | Prefill tok/s | TTFT ms | VRAM GiB | Source |
|---|---|---|---|---|---|---|---|---|---|---|---|
| r_m6j1zxwvp-f | 2026-10-16 | chat-short | Q4_K_M | llama.cpp · 8931 (9725a313b) | 112 | 256 | 160.6 | 464.3 | 241 | – | Measured by llm-speed@llmspeed |
| r_m6j1zxwvp-f | 2026-10-16 | chat-long | Q4_K_M | llama.cpp · 8931 (9725a313b) | 3182 | 1024 | 160.4 | 4588.8 | 693 | – | Measured by llm-speed@llmspeed |
| r_m6j1zxwvp-f | 2026-10-16 | agent-trace | Q4_K_M | llama.cpp · 8931 (9725a313b) | 3454 | 840 | 162.7 | 6547.5 | 330 | – | Measured by llm-speed@llmspeed |
| r_m6j1zxwvp-f | 2026-10-16 | concurrent-decode | Q4_K_M | llama.cpp · 8931 (9725a313b) | 1024 | 256 | 163.5 | – | – | – | Measured by llm-speed@llmspeed |
Questions
Can an RTX 4090 run Qwen3.6-35B-A3B?
Yes: an RTX 4090 runs Qwen3.6 35B-A3B at 160.6 tok/s at Q4_K_M, measured on llama.cpp with a 112-token prompt in October 2026.
How many tokens per second does Qwen3.6-35B-A3B get on an RTX 4090?
Decode speed measured by llm-speed, by quantization and runtime: Q4_K_M on llama.cpp: 160.6 tok/s. Based on 1 run measured by llm-speed; single-request chat-short unless noted.
What is the fastest quant of Qwen3.6-35B-A3B that fits on an RTX 4090?
Q4_K_M on llama.cpp was the fastest measured setting at 160.6 tok/s; it fit because it ran. Other quants are listed only where a run measured them.
How much VRAM does Qwen3.6-35B-A3B use on an RTX 4090?
Memory use was not recorded for these runs. The VRAM calculator gives an estimate, not a measurement.