Can I Run Devstral on NVIDIA GeForce RTX 5090?
作者： Jakub Rusinowski · 最后更新： 2026年9月11日
Yes, comfortably — you'll have ~17.5 GB of headroom running Devstral Small 2505 24B at Q4_K_M (14.5 GB, ~82 tok/s (est.)) with room for up to 64K context.
See what else this hardware can run →
作为亚马逊联盟成员，我们从符合条件的购买中获得收入。云 GPU 链接为推荐链接——我们可能获得佣金，您无需额外付费。
NVIDIA GeForce RTX 5090 Specs
| VRAM | 32 GB |
| Memory Bandwidth | 1792 GB/s |
Devstral Small 2505 24B on the NVIDIA GeForce RTX 5090: VRAM by quantization
| Quant | VRAM needed | Fits 32 GB? | Max context | Whole PC |
|---|---|---|---|---|
| F16 | 49.5 GB | ✗ No | — | — |
| Q8_0 | 27 GB | ✓ Yes | 32K | — |
| Q6_K | 21.2 GB | ✓ Yes | 64K | — |
| Q5_K_M | 18.5 GB | ✓ Yes | 64K | — |
| Q4_K_M | 16 GB | ✓ Yes | 64K | — |
| Q3_K_M | 11.7 GB | ✓ Yes | 64K | — |
| Q2_K | 9.4 GB | ✓ Yes | 128K | — |
Assumes a 4K-token context with an f16 KV cache. A longer window needs more; a quantized KV cache needs less. “Max context” is the largest window that still fits in 32 GB. Figures are estimates from parameter count, quantization and memory bandwidth — the analyzer lets you tune KV-cache quant and context.
Context costs VRAM too. At Q4_K_M on the NVIDIA GeForce RTX 5090: 8K 16.6 GB · 32K 20.7 GB · 128K 36.8 GB ✗. Past 128K it no longer fits 32 GB — a q8_0 KV cache buys roughly half of that back, which the calculator will price for you.
No compatible GPU? Devstral Small 2505 24B on 24 GB of system RAM, CPU only: runs comfortably.
Devstral Sizes That Fit the NVIDIA GeForce RTX 5090
| Devstral Small 2505 24B | Q4_K_M · 14.5 GB · ~82 tok/s (est.) |
| Devstral Small 2 24B | Q4_K_M · 14.5 GB · ~81 tok/s (est.) |
| Devstral 2 22B | Q4_K_M · 13.3 GB · ~87 tok/s (est.) |
At 2 hrs/day, buying (~$1,999) beats renting at $0.34/hr after about 8.2 years.
Affiliate links — we may earn a commission if you sign up, at no extra cost to you.
Cloud rates verified 2026-07 — estimates, and marketplace prices vary. Buying price is GPU MSRP only, not a full PC.
FAQ
Does the NVIDIA GeForce RTX 5090 have enough VRAM for Devstral?
Yes, comfortably — you'll have ~17.5 GB of headroom running Devstral Small 2505 24B at Q4_K_M (14.5 GB, ~82 tok/s (est.)) with room for up to 64K context.
Which quantization of Devstral should I use on the NVIDIA GeForce RTX 5090?
Devstral Small 2505 24B at Q4_K_M quantization (14.5 GB), estimated ~82 tokens/sec, up to 64K context.
Every Devstral size on the NVIDIA GeForce RTX 5090
| Size | VRAM | Verdict | Speed |
|---|---|---|---|
| Devstral-2 123B | 74.3 GB | Does not fit | — |
| Devstral Small 2505 24B | 14.5 GB | Runs | ~82 tok/s |
| Devstral Small 2 24B | 14.5 GB | Runs | ~81 tok/s |
| Devstral 2 22B | 13.3 GB | Runs | ~87 tok/s |
Devstral on Other GPUs
- Devstral on Apple M4
- Devstral on Apple M2 Pro
- Devstral on Apple M1 Pro
- Devstral on Apple M5
- Devstral on Apple M3 Pro
Popular Models on the NVIDIA GeForce RTX 5090
- DeepSeek R1 on NVIDIA GeForce RTX 5090
- Qwen 3 on NVIDIA GeForce RTX 5090
- Qwen 2.5 Family on NVIDIA GeForce RTX 5090
VRAM Tier
Troubleshooting
Buying Guide
← Can I Run It? | Devstral Model Page | NVIDIA GeForce RTX 5090 GPU Page | Check Your Hardware