RX 7900 XTX 24GBBest value
C31 tok/s decode
~$999 MSRP
Can it run?
Too heavy
Using Q4_K_M in Ollama
Fit status
Too heavy
Decode
33.0 tok/s
TTFT
5858 ms
Safe context
9K
Memory
27.3 GB / 16.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | F | Too heavy | 33.0 tok/s | 8521 ms | 16K |
| Chat | F | Too heavy | 33.0 tok/s | 3195 ms | 5K |
| Coding | F | Too heavy | 33.0 tok/s | 5858 ms | 9K |
| RAG | F | Too heavy | 33.0 tok/s | 10651 ms | 16K |
| Reasoning | F | Too heavy | 33.0 tok/s | 6923 ms | 9K |
How Qwen 2.5 Coder 32B (32B params) fits at each quantization level on RTX 5080 Laptop 16GB (16.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 12.5 GB | Low | C44 |
Q3_K_S | 3 | 15.7 GB | Low | C44 |
NVFP4 | 4 | 17.9 GB | Medium | F0 |
Q4_K_M | 4 | 19.5 GB | Medium | F0 |
Q5_K_M | 5 | 23.0 GB | High | F0 |
Q6_K | 6 | 26.2 GB | High | F0 |
Q8_0 | 8 | 34.2 GB | Very High | F0 |
F16 | 16 | 65.6 GB | Maximum | F0 |
Upgrade options
~$999 MSRP
~$1,099 MSRP
NVIDIA A100 40GBBiggest leap~$10,000 MSRP
RTX PRO 5000 Blackwell 48GBNVIDIA upgrade