RX 7900 XTX 24GBBest value
C99.4 tok/s decode
~$999 MSRP
Can it run?
Too heavy
Using Q4_K_M in vLLM
Fit status
Too heavy
Decode
32.3 tok/s
TTFT
5993 ms
Safe context
8K
Memory
22.7 GB / 12.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | F | Too heavy | 32.3 tok/s | 8717 ms | 17K |
| Chat | F | Too heavy | 32.3 tok/s | 3269 ms | 4K |
| Coding | F | Too heavy | 32.3 tok/s | 5993 ms | 8K |
| RAG | F | Too heavy | 32.3 tok/s | 10897 ms | 17K |
| Reasoning | F | Too heavy | 32.3 tok/s | 7083 ms | 8K |
How Qwen3-VL 30B A3B Instruct (30B params) fits at each quantization level on RTX A2000 12GB (12.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 11.7 GB | Low | C45 |
Q3_K_S | 3 | 14.7 GB | Low | F0 |
NVFP4 | 4 | 16.8 GB | Medium | F0 |
Q4_K_M | 4 | 18.3 GB | Medium | F0 |
Q5_K_M | 5 | 21.6 GB | High | F0 |
Q6_K | 6 | 24.6 GB | High | F0 |
Q8_0 | 8 | 32.1 GB | Very High | F0 |
F16 | 16 | 61.5 GB | Maximum | F0 |
Upgrade options
~$999 MSRP
~$1,099 MSRP
RTX 5090 32GBNVIDIA upgrade~$1,999 MSRP