RX 7900 XT 20GBBudget pick
C112.4 tok/s decode
~$899 MSRP
Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
15.9 tok/s
TTFT
12157 ms
Safe context
30K
Memory
9.2 GB / 17.3 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 15.9 tok/s | 17683 ms | 54K |
| Chat | C | Runs well | 15.9 tok/s | 6631 ms | 16K |
| Coding | C | Runs well | 15.9 tok/s | 12157 ms | 30K |
| RAG | C | Runs well | 15.9 tok/s | 22104 ms | 54K |
| Reasoning | C | Runs well | 15.9 tok/s | 14367 ms | 30K |
How Mistral 7B Instruct v0.2 (7B params) fits at each quantization level on MacBook Air M3 24GB (17.3 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 2.7 GB | Low | D33 |
Q3_K_S | 3 | 3.4 GB | Low | D34 |
NVFP4 | 4 | 3.9 GB | Medium | D34 |
Q4_K_M | 4 | 4.3 GB | Medium | D35 |
Q5_K_M | 5 | 5.0 GB | High | D36 |
Q6_K | 6 | 5.7 GB | High | D37 |
Q8_0Best for your GPU | 8 | 7.5 GB | Very High | D39 |
F16 | 16 | 14.3 GB | Maximum | C44 |
Upgrade options
~$899 MSRP
~$999 MSRP
~$1,999 MSRP
RTX A4500 20GBBiggest leap