Can I Run Devstral on Apple M3?
作者： Jakub Rusinowski · 最后更新： 2026年9月11日
Yes, but it's tight — Devstral Small 2505 24B at Q4_K_M needs 16.6 GB at 8K context (14.5 GB weights + 2.1 GB KV cache/overhead), leaving ~1.4 GB of the Apple M3's 18 GB usable memory, at ~4 tok/s (est.), with room for up to 16K context.
See what else this hardware can run →
或在 Vast.ai 比较，低至 $0.35/小时 (typical low · varies)
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。
Apple M3 Specs
| VRAM | 24 GB unified memory |
| Memory Bandwidth | 100 GB/s |
Devstral Small 2505 24B on the Apple M3: VRAM by quantization
| Quant | VRAM needed | Fits 24 GB? | Max context | Whole PC |
|---|---|---|---|---|
| F16 | 50.1 GB | ✗ No | — | — |
| Q8_0 | 27.6 GB | ✗ No | — | — |
| Q6_K | 21.8 GB | ✗ No | — | — |
| Q5_K_M | 19.2 GB | ✗ No | — | — |
| Q4_K_M | 16.6 GB | ✓ Yes | 16K | — |
| Q3_K_M | 12.4 GB | ✓ Yes | 32K | — |
| Q2_K | 10 GB | ✓ Yes | 32K | — |
Assumes an 8K-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 24 GB of unified memory. 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 Apple M3: 8K 16.6 GB · 32K 20.7 GB ✗ · 128K 36.8 GB ✗. Past 32K it no longer fits 24 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 16 GB of system RAM, CPU only: runs, but it is tight.
Devstral Sizes That Fit the Apple M3 (at 8K context)
| Devstral Small 2505 24B | Q4_K_M · 16.6 GB at 8K context · ~4 tok/s (est.) |
| Devstral Small 2 24B | Q4_K_M · 16.6 GB at 8K context · ~4 tok/s (est.) |
Devstral needs ~17 GB but this GPU has 24 GB. Rent a RTX 4090 (24 GB)-class GPU by the hour instead of buying one:
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
Can I run Devstral on the Apple M3?
Yes, but it's tight — Devstral Small 2505 24B at Q4_K_M needs 16.6 GB at 8K context (14.5 GB weights + 2.1 GB KV cache/overhead), leaving ~1.4 GB of the Apple M3's 18 GB usable memory, at ~4 tok/s (est.), with room for up to 16K context.
What's the best Devstral size for the Apple M3?
Devstral Small 2505 24B at Q4_K_M quantization needs 16.6 GB at 8K context (14.5 GB weights + 2.1 GB KV cache/overhead), estimated ~4 tokens/sec, up to 16K context.
Every Devstral size on the Apple M3
| Size | VRAM at 8K | Verdict | Speed |
|---|---|---|---|
| Devstral-2 123B | 77.9 GB | Does not fit | — |
| Devstral Small 2505 24B | 16.6 GB | Runs | ~4 tok/s |
| Devstral Small 2 24B | 16.6 GB | Runs | ~4 tok/s |
Devstral on Other GPUs
- Devstral on NVIDIA GeForce RTX 4090
- Devstral on NVIDIA GeForce RTX 3090
- Devstral on AMD Radeon RX 7900 XTX
- Devstral on Apple M4 Pro
- Devstral on Apple M2
VRAM Tier
Troubleshooting
- Ollama: "model requires more system memory than is available"
- Which GGUF quant should I download? (Q4 vs Q5 vs Q8)
Buying Guide
← Can I Run It? | Devstral Model Page | Apple M3 GPU Page | VRAM calculator | Check Your Hardware