Magistral Small 2507 needs ~14.2 GB VRAM. RTX 5070 12GB has 12.0 GB. With Q2_K quantization, expect ~22 tok/s.
Operating mode
Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.
Current mode
Balanced
Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.
Select quantization to explore
7.5 GB over capacity — needs offload or smaller quantization
Fit status
Too heavy
Decode
8.8 tok/s
TTFT
21884 ms
Safe context
4K
Memory
19.5 GB / 12.0 GB
Offload
40%
It fits through host-memory offload, and offload is the main reason performance drops.
CPU or host-memory offload is active
About 20% of the working set spills out of accelerator memory, which usually hurts latency and sustained decode throughput.
Very little memory headroom
You can run the model, but there is not much room left for longer context, bigger batches, extra apps, or future model updates.
Remove offload with more accelerator memory
Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.
Buy headroom, not only minimum fit
A slightly larger memory tier gives you safer context growth and makes the recommendation more future-proof.
Increase host RAM if you keep offloading
This setup may need roughly 1.5 GB of extra host RAM just for the offloaded portion, before OS and other tools.
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | F | Too heavy | 10.1 tok/s | 10461 ms | 4K |
| Coding | F | Too heavy | 8.8 tok/s | 21884 ms | 4K |
| Agentic Coding | F | Too heavy | 7.0 tok/s | 40500 ms | 4K |
| Reasoning | F | Too heavy | 8.8 tok/s | 25863 ms | 4K |
| RAG | F | Too heavy | 7.0 tok/s | 50626 ms | 4K |
How Magistral Small 2507 (24B params) fits at each quantization level on RTX 5070 12GB (12.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 9.4 GB | Low | F0 |
Q3_K_S | 3 | 11.8 GB | Low | F0 |
NVFP4 | 4 |
Copy-paste commands to run Magistral Small 2507 on your machine.
Run
ollama run magistralUpgrade options
RTX 4000 Ada 20GBBudget pickMakes the model fit on the accelerator instead of staying completely out of reach.
Removes host-memory offload, which is usually the single biggest latency and throughput win.
~$1,250 MSRP
RTX 3090 24GBBest valueMakes the model fit on the accelerator instead of staying completely out of reach.
Removes host-memory offload, which is usually the single biggest latency and throughput win.
~$1,499 MSRP
RTX 4090 24GBNVIDIA upgradeMakes the model fit on the accelerator instead of staying completely out of reach.
Removes host-memory offload, which is usually the single biggest latency and throughput win.
~$1,599 MSRP
Yes, RTX 5070 12GB can run Magistral Small 2507 at Q2_K quantization (Very compromised (needs ~1.5 GB host RAM)). The recommended Q4_K_M requires 19.5 GB which exceeds available memory, but at Q2_K it needs only 14.2 GB. Expected decode speed: 22.4 tok/s.
Magistral Small 2507 (24B parameters) requires approximately 19.5 GB at Q4_K_M quantization. On RTX 5070 12GB, it fits at Q2_K using 14.2 GB.
The recommended quantization is Q4_K_M, but on RTX 5070 12GB the best fitting quantization is Q2_K, which uses 14.2 GB.
On RTX 5070 12GB, Magistral Small 2507 achieves approximately 22.4 tokens per second decode speed with a time-to-first-token of 8637ms using Q2_K quantization.
For coding workloads, Magistral Small 2507 on RTX 5070 12GB receives a F grade with 8.8 tok/s and 4K context.
On RTX 5070 12GB, Magistral Small 2507 can safely use up to 4K tokens of context at Q2_K quantization. The model's official context limit is 131K, but available memory constrains the safe maximum.
Paste this snippet into any page to show a live fit card.
<iframe src="https://willitrunai.com/embed/magistral-small-2507-on-rtx-5070-12gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>
Preview:
| Medium |
| F0 |
Q4_K_M | 4 | 14.6 GB | Medium | F0 |
Q5_K_M | 5 | 17.3 GB | High | F0 |
Q6_K | 6 | 19.7 GB | High | F0 |
Q8_0 | 8 | 25.7 GB | Very High | F0 |
F16 | 16 | 49.2 GB | Maximum | F0 |
Remove offload with more accelerator memory. Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.