Can an RTX 4090 run Qwen3-32B?
llm-speed has not measured Qwen3-32B on RTX 4090 itself yet, so there is no llm-speed verdict. 1 community-submitted run exists (fastest 43.5 tok/s at Q4_K_M on llama.cpp); community-submitted, not measured by llm-speed, so it gives no verdict.
1 signed run on llm-speed. Single-request decode speed in tokens per second; the short-chat workload unless noted. This run is community-submitted (signed on the submitter's own machine, not measured by llm-speed, and not used for llm-speed verdicts). Verdicts and headline numbers use only llm-speed's own runs and fit tests.
Measured tok/s by quant and runtime
| Quant | Runtime | Build | Decode tok/s (best run) | Workload | Runs | Source |
|---|---|---|---|---|---|---|
| Q4_K_M | llama.cpp | commit df18bee | 43.5 | chat-short | 1 | Community-submitted |
llm-speed has not measured this pair itself, so there is no llm-speed verdict or fastest quant. The rows above are community-submitted, not measured by llm-speed.
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_ayy9rjgha5r | 2026-10-14 | chat-short | Q4_K_M | llama.cpp · commit df18bee | 110 | 256 | 43.5 | 778.6 | 141 | – | Community-submitted |
| r_ayy9rjgha5r | 2026-10-14 | chat-long | Q4_K_M | llama.cpp · commit df18bee | 3147 | 1024 | 40.7 | 2333.3 | 1349 | – | Community-submitted |
| r_ayy9rjgha5r | 2026-10-14 | agent-trace | Q4_K_M | llama.cpp · commit df18bee | 3247 | 840 | 41.9 | 5982.3 | 342 | – | Community-submitted |
| r_ayy9rjgha5r | 2026-10-14 | concurrent-decode | Q4_K_M | llama.cpp · commit df18bee | 1024 | 256 | 42.7 | – | – | – | Community-submitted |
Questions
Can an RTX 4090 run Qwen3-32B?
llm-speed has not measured Qwen3-32B on RTX 4090 itself yet, so there is no llm-speed verdict. 1 community-submitted run exists (fastest 43.5 tok/s at Q4_K_M on llama.cpp); community-submitted, not measured by llm-speed, so it gives no verdict.
What did community runs measure for Qwen3-32B on an RTX 4090?
1 community-submitted run (community-submitted, not measured by llm-speed): Q4_K_M on llama.cpp: 43.5 tok/s. They are listed for reference and give no llm-speed verdict.