Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
138.2 tok/s
TTFT
1400 ms
Safe context
26K
Memory
2.5 GB / 4.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | B | Runs well | 149.8 tok/s | 1880 ms | 52K |
| Chat | B | Runs well | 149.8 tok/s | 705 ms | 13K |
| Coding | B | Runs well | 138.2 tok/s | 1400 ms | 26K |
| RAG | B | Runs well | 149.8 tok/s | 2350 ms | 52K |
| Reasoning | B | Runs well | 149.8 tok/s | 1528 ms | 26K |
How Falcon H1 Tiny 90M Instruct (0.09000000357627869B params) fits at each quantization level on RTX 3050 Ti Laptop 4GB (4.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 0.0 GB | Low | D29 |
Q3_K_S | 3 | 0.0 GB | Low | D29 |
NVFP4 | 4 |
0.1 GB |
| Medium |
| D29 |
Q4_K_M | 4 | 0.1 GB | Medium | D29 |
Q5_K_M | 5 | 0.1 GB | High | D29 |
Q6_K | 6 | 0.1 GB | High | D29 |
Q8_0 | 8 | 0.1 GB | Very High | D29 |
F16Best for your GPU | 16 | 0.2 GB | Maximum | D30 |
