Radeon PRO W7900
AMD
ExcellentRuns Mistral Small 4 119B Q4_K_M entirely in memory
Model · Mistral AI
Mistral Small 4 119B: VRAM and hardware requirements
At Q4_K_M with an 8K context, Mistral Small 4 119B needs about 72.5 GB: a GPU with 80 GB or more runs it entirely in VRAM.
- Parameters119.4B
- Active6.5B
- Max context256K
- LicenseApache-2.0
Specifications
- Organization
- Mistral AI
- Architecture
- mistral4
- Released
- January 2026
- Layers
- 36
- Hidden size
- 4096
- Attention heads
- 32
- KV heads
- MLA (rank 256)
- Head dimension
- 128
- Vocabulary
- 131,072
- KV cache per 1K tokens (f16)
- 0.022 GB
- Repository
- mistralai/Mistral-Small-4-119B-2603
Official specData updated todaySource: Hugging Face config.json + model card
Quantizations and memory
| Quantization | File size | Memory needed | Minimum GPU memory | Quality |
|---|---|---|---|---|
| BF16 | Not available | ~235 GB | – | Reference (unquantized) |
| Q8_0 | 118 GB | ~125 GB | – | Near-lossless |
| Q6_K | 95.7 GB | ~96.9 GB | – | Near-lossless |
| Q5_K_M | 79.2 GB | ~84.2 GB | 96 GB | Very good |
| Q4_K_MMost used | 67.6 GB | ~72.5 GB | 80 GB | Good — the usual choice |
| IQ4_XS | 59.4 GB | ~66.1 GB | 80 GB | Good — the usual choice |
| Q3_K_M | 50.8 GB | ~59.4 GB | 80 GB | Noticeable loss |
| Q2_K | 39.2 GB | ~47.3 GB | 48 GB | Significant loss |
File sizes: not available until imported from the model repository. Memory = weights (estimated from bits per weight) + KV cache + runtime overhead + safety margin. (BF16 ≈ 16 bpw, Q8_0 ≈ 8.5 bpw, Q6_K ≈ 6.57 bpw, Q5_K_M ≈ 5.7 bpw, Q4_K_M ≈ 4.9 bpw, IQ4_XS ≈ 4.46 bpw, Q3_K_M ≈ 4 bpw, Q2_K ≈ 3.17 bpw)
Q4, Q5 or Q8: how to choose a quantization
GPUs that run Mistral Small 4 119B
| RTX 509032 GB · 64 GB RAM~72 GB · 8–16 tok/s estimate | Possible but slow |
|---|---|
| RTX 409024 GB · 64 GB RAM~72 GB · 7–14 tok/s estimate | Possible but slow |
| RTX 309024 GB · 64 GB RAM~72 GB · 7–14 tok/s estimate | Possible but slow |
| RX 7900 XTX24 GB · 64 GB RAM~72.1 GB · 6.5–13 tok/s estimate | Possible but slow |
| RTX 5060 Ti 16 GB16 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RTX 5070 Ti16 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RTX 508016 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RTX 4060 Ti 16 GB16 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RTX 4070 Ti SUPER16 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RX 9070 XT16 GB · 64 GB RAM~72.1 GB · No data | Incompatible |
| RTX 3060 12 GB12 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RTX 407012 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RTX 507012 GB · 64 GB RAM~72 GB · No data | Incompatible |
| RTX 40608 GB · 64 GB RAM~72 GB · No data | Incompatible |
Macs that run Mistral Small 4 119B
| M3 Ultra (80-core GPU) · 512 GBMac Studio~71.8 GB · 59–99 tok/s estimate | Excellent |
|---|---|
| M4 Max (40-core GPU) · 128 GBMacBook Pro 14, MacBook Pro 16~71.8 GB · 47–76 tok/s estimate | Excellent |
| M3 Max (40-core GPU) · 128 GBMacBook Pro 14, MacBook Pro 16~71.8 GB · 39–61 tok/s estimate | Excellent |
| M4 Pro (20-core GPU) · 64 GBMac mini, MacBook Pro 14~71.8 GB · No data | Incompatible |
| M4 Max (32-core GPU) · 36 GBMacBook Pro 14, MacBook Pro 16~71.8 GB · No data | Incompatible |
| M4 · 32 GBMacBook Air, MacBook Pro 14~71.8 GB · No data | Incompatible |
| M5 · 32 GBMacBook Pro 14~71.8 GB · No data | Incompatible |
| M3 · 24 GBMacBook Air, MacBook Pro 14~71.8 GB · No data | Incompatible |
EstimateData updated 40 days agoSource: Run LLM Locally engine
Limits in plain words
Mistral Small 4 119B is a mixture-of-experts model: all 119.4B parameters must be in memory, but only 6.5B are read per token, so it generates much faster than a dense model of the same size.
The KV cache grows with context: at 32K it adds about 0.7 GB on top of the weights.
With under 8 GB of VRAM, use a smaller quantization or a smaller model of the same family.
Where to buy
Prices and links follow the market you choose.
NVIDIA DGX Spark (GB10)
Unified memory PC
ExcellentRuns Mistral Small 4 119B Q4_K_M entirely in memory
No current offer imported for this market: the link opens a marketplace search.
Frequently asked questions
How much VRAM does Mistral Small 4 119B need?
About 72.5 GB at Q4_K_M with an 8K context, including the KV cache and runtime overhead. At Q8_0 it needs about 125 GB. A longer context adds more: about 0.7 GB of KV cache at 32K.
Can Mistral Small 4 119B run on a Mac?
Yes on a Mac whose GPU can use at least 71.8 GB, e.g. M4 Max (40-core GPU) 128 GB. macOS lets the GPU use about two thirds to three quarters of unified memory by default.
How fast is Mistral Small 4 119B?
On a RTX 4090, we estimate 7–14 tok/s at Q4_K_M. This is an estimate from memory bandwidth (confidence low) until benchmarks are imported.