Yes — with 68.4 GB to spare
Ministral 3 3B at Q4_K_M fits your M2 Max · 96 GB entirely in unified memory at 8K context, at an estimated 103 tokens per second. There is room for its full 256K window.
Fully in unified memory
8K context
Q4_K_M · 2.2 GB
Apache 2.0
Released Dec 2025
Vision
Edge model with a vision encoder and a 256K window. Apache 2.0.
Hardware requirements What hardware do I need to run Ministral 3 3B? →At Q4_K_M with an 8K context: a 6 GB GPU (RTX 3060 Laptop) or a 8 GB Mac (M3). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.
The VRAM budget
weights 2.2 GB
Weights 2.2 GB
KV cache @ 8K 0.81 GB
Runtime overhead 0.6 GB
Free 68.4 GB of 72.0 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| F16 | 7.2 GB | 8.6 GB | 256K | 31 | Reference | Long context |
| Q8_0 | 3.8 GB | 5.2 GB | 256K | 59 | −0.1% ppl | Long context |
| Q6_K | 2.9 GB | 4.4 GB | 256K | 76 | −0.4% ppl | Long context |
| Q5_K_M | 2.5 GB | 4.0 GB | 256K | 88 | −0.8% ppl | Long context |
| Q4_K_M | 2.2 GB | 3.6 GB | 256K | 103 | −1.9% ppl | Recommended |
| Q3_K_M | 1.8 GB | 3.2 GB | 256K | 128 | −5.4% ppl | Long context |
| Q2_K | 1.5 GB | 2.9 GB | 256K | 149 | −15% ppl | Long context |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.
How to run it
$ pip install mlx-lm $ mlx_lm.generate --model mlx-community/Ministral-3-3B-Instruct-2512-4bit \ --max-tokens 512 --prompt "Hello"
Apple's own array framework. The fastest path on Apple Silicon. More on MLX.
01Download is 2.2 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02macOS caps what the GPU may wire down at about 72.0 GB of 96 GB. Raising it with iogpu.wired_limit_mb is possible, and risky.
03There is room to go to the model's full 256K context on this card.
Related models on an M2 Max · 96 GB
| Model | Total | Tok/s est. | |
|---|---|---|---|
| Ministral 3 8B 8.92B · same family |
Q4_K_M · 6.7 GB | 45 | Runs great |
| Ministral 3 14B 13.9B · same family |
Q4_K_M · 9.7 GB | 29 | Runs great |
| Phi-4-mini 3.8B 3.84B · similar size |
Q4_K_M · 3.8 GB | 104 | Runs great |
| Qwen3 4B 4.02B · similar size |
Q4_K_M · 4.0 GB | 99 | Runs great |
| Gemma 3 4B 4.3B · similar size |
Q4_K_M · 3.3 GB | 92 | Runs great |
| Granite 4.1 3B 3.4B · similar size |
Q4_K_M · 3.1 GB | 117 | Runs great |
| Llama 3.2 3B Instruct 3.21B · similar size |
Q4_K_M · 3.3 GB | 124 | Runs great |
Ministral 3 3B on other hardware
| Device | Total | Tok/s est. | |
|---|---|---|---|
| GeForce RTX 3060 Laptop 5.3 GB usable · 336 GB/s |
3.6 GB | 94 | Runs great |
| GeForce RTX 4060 Laptop 7.0 GB usable · 256 GB/s |
3.6 GB | 72 | Runs great |
| M3 · 8 GB 5.4 GB usable · 100 GB/s |
3.6 GB | 26 | Runs great |
| GeForce RTX 3060 12 GB 10.6 GB usable · 360 GB/s |
3.6 GB | 101 | Runs great |
| M4 · 16 GB 10.7 GB usable · 120 GB/s |
3.6 GB | 31 | Runs great |
| GeForce RTX 4090 22.4 GB usable · 1008 GB/s |
3.6 GB | 282 | Runs great |
| Radeon RX 7900 XTX 22.4 GB usable · 960 GB/s |
3.6 GB | 268 | Runs great |
| M3 Max · 64 GB 48.0 GB usable · 400 GB/s |
3.6 GB | 103 | Runs great |