What it costs to run Devstral Small 2 24B
Trying to decide what to buy? The best GPU for Devstral Small 2 24B →
Cheapest way to rent it right now
$0.03/hr
NVIDIA L4 · Q6_K · spot · ≈ $8/mo at 8h/day
Live price captured Sat, 07 Nov 2026 02:20:13 GMT.Referral link — we may earn a commission at no cost to you; it never changes which card is cheapest.
Memory needed, by quantisation
At 16k context. Weights + KV cache + ~0.8 GB overhead.
| Quant | Weights | KV cache | Total | Quality |
|---|---|---|---|---|
| F16 | 44.7 GB | 2.5 GB | 48.0 GB | Full precision. Reference quality, twice the size of Q8 for no practical gain in most chat use. |
| Q8_0 | 23.8 GB | 2.5 GB | 27.0 GB | Effectively lossless. Use when VRAM is not the constraint. |
| Q6_K | 18.3 GB | 2.5 GB | 21.6 GB | Very close to Q8 at meaningfully less memory. A safe high-quality pick. |
| Q5_K_M | 15.9 GB | 2.5 GB | 19.2 GB | Small, hard-to-notice quality loss. Good balance. |
| Q5_0 | 15.5 GB | 2.5 GB | 18.8 GB | Older-style 5-bit. Q5_K_M is usually the better pick at the same size. |
| Q4_K_M | 13.6 GB | 2.5 GB | 16.8 GB | The community default. Best quality-per-gigabyte for most people. |
| Q4_0 | 12.7 GB | 2.5 GB | 16.0 GB | Older-style 4-bit, measurably worse than Q4_K_M at a similar size. Avoid unless required. |
| IQ4_XS | 11.9 GB | 2.5 GB | 15.2 GB | Newer 4-bit, smaller than Q4_K_M with comparable quality. Needs a recent llama.cpp. |
| Q3_K_M | 10.9 GB | 2.5 GB | 14.2 GB | Noticeable degradation. Use to fit a larger model that would otherwise not run. |
| Q2_K | 9.4 GB | 2.5 GB | 12.6 GB | Heavy degradation. Almost always better to run a smaller model at Q4_K_M instead. |
| IQ3_XS | 9.2 GB | 2.5 GB | 12.5 GB | Aggressive. Usually better than a smaller model at Q4, but test before trusting it. |
Cards that can run it
Best quantisation each card fits at 16k context, with live rental price where we track one.
How these numbers are produced
- Model shape comes from the model's own config.json — 40 layers, 32 attention heads, 8 KV heads.
- Weight size is parameters × effective bits-per-weight. Those constants are checked against real published quantised file sizes — median error 0.7% across 40 measurements.
- KV cache is 2 × layers × KV-heads × head-dim × context × 2 bytes. Using KV-heads rather than attention heads is what makes this correct for grouped-query attention; treating a GQA model as multi-head overstates the cache by up to 8×.
- Tokens/sec is an estimate, not a benchmark. Generation is memory-bandwidth-bound, so this is bandwidth ÷ weight-bytes derated to 75%. Real throughput depends on your runtime, batch size and kernels. We don't run our own hardware tests — see our editorial policy.
- Some memory-bandwidth figures are not yet independently verified.Where that's the case we show no tokens/sec at all rather than a number we can't stand behind.
- Rental prices are pulled hourly from provider APIs (last updated Sat, 07 Nov 2026 02:20:13 GMT). Spot/interruptible pricing can change or vanish without notice.
Try it interactively
Drag the context slider and watch the KV cache fill the card
Worth reading before you buy
How much VRAM do you need?
About 0.57 GB per billion parameters at 4-bit, plus room for context: the rule behind the numbers on this page, with worked examples.
The best GPU for local AI right now
The card with enough VRAM for your model at the lowest price, budget by budget.
Rent a cloud GPU or buy one?
Rent, unless you'll pass about 2,000 hours of use: the break-even and the costs people forget on both sides.
What quantization means
Why a 4-bit copy of a model fits on a normal PC, and what you give up for it.