Mixtral 34Bx2 MoE 60B — Hardware Requirements & GPU Compatibility
Chat
Mixtral 34Bx2 MoE 60B is a 60.8B-parameter open language model from cloudyu in the Mixtral family. It supports a context window of up to 200,000 tokens. At Q4_K_M it needs about 37.29 GB of VRAM — see which GPUs and Macs can run it below.
Doesn't fit a 24 GB GPU
Smallest consumer fit → RTX 5090 · Q2_K · 30.2 GB · ~44 tok/s
Macs from 36 GB (Q2_K · ~12 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Mixtral 34Bx2 MoE 60B
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve cloudyu/Mixtral_34Bx2_MoE_60B --max-model-len 16384vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on H100 (80 GB) (39 GB headroom) · Q4_K_M
- Generation speed
- ~58 tok/s
- generation speed
- Cost per 1M output tokens
- $12.79
- per 1M output tokens
How Much VRAM Does Mixtral 34Bx2 MoE 60B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 26.6 GB | 34.2 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Does not fit: needs 30.2 GB | |
| 30.4 GB | 38 GB | 3-bit medium — 3-bit medium quantization | Does not fit: needs 34 GB | |
| 37.3 GB | 44.8 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Does not fit: needs 40.8 GB | |
| 44.1 GB | 51.7 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Does not fit: needs 47.7 GB | |
| 51.0 GB | 58.5 GB | 6-bit, very good — 6-bit quantization, very good quality | Does not fit: needs 54.5 GB | |
| 61.6 GB | 69.2 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Does not fit: needs 65.1 GB | |
| 122.4 GB | 130 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 126 GB |
Selected Q4_K_M
4-bit medium quantization40.8GBat 16K context
- Weights
- 36.5 GB
- KV cache
- 4.0 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 16.8 GB more
Or .
Which Hardware Can Run Mixtral 34Bx2 MoE 60B?
Q4_K_M · 37.3 GBMixtral 34Bx2 MoE 60B (Q4_K_M) requires 37.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 49+ GB is recommended. Using the full 200K context window can add up to 48.7 GB, bringing total usage to 85.9 GB. No consumer GPU has enough memory. 67 devices with unified memory can run Mixtral 34Bx2 MoE 60B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M5 Max (48 GB).
32–40 GB
Q4_K_M · best fit Q2_K (30.2 GB)
1 GPU · 3 Mac classes · 1 device family
RTX 5090~44Simulate output on RTX 5090M5 Max36 GB~12Simulate output on Mac Studio M5 Max (36 GB)M532 GB~4Simulate output on MacBook Air 13" M5 (32 GB)Orin~5Simulate output on NVIDIA Jetson AGX Orin 32GB48 GB
48 GB tier →fits · 7.2 GB free · up to 32K ctx
2 Mac classes · 1 device family
M5 Max48 GB~12Simulate output on Mac Studio M5 Max (48 GB)M5 Pro48 GB~6Simulate output on Mac Mini M5 Pro (48 GB)Snapdragon X2 Elite Extreme (X2E-96-100)~4Simulate output on Snapdragon X2 Elite Extreme Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to Mixtral 34Bx2 MoE 60B
Mixtral family · 7 models →Similar models
1| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen2 57B A14B InstructAlibaba | — | 35.7 GB | —needs > 24 GB | same size |
Specifications
- Publisher
- cloudyu
- Family
- Mixtral
- Parameters
- 60.8B
- Architecture
- MixtralForCausalLM
- Context Length
- 200,000 tokens
- Vocabulary Size
- 64,000
- Release Date
- 2024-01-05
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Mixtral 34Bx2 MoE 60B need?
Mixtral 34Bx2 MoE 60B requires 37.3 GB of VRAM at Q4_K_M, or 122.4 GB at BF16. Full 200K context adds up to 48.6 GB (85.9 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 60.8B × 4.8 bits ÷ 8 = 36.5 GB
KV Cache + Overhead ≈ 0.8 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 49.4 GB (at full 200K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Mixtral 34Bx2 MoE 60B?
Yes, at Q3_K_M (30.4 GB) or lower. Higher quantizations like Q4_K_M (37.3 GB) exceed the NVIDIA GeForce RTX 5090's 32 GB.
What's the best quantization for Mixtral 34Bx2 MoE 60B?
For Mixtral 34Bx2 MoE 60B, Q4_K_M (37.3 GB) offers the best balance of quality and VRAM usage. Q5_K_M (44.1 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 26.6 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Mixtral 34Bx2 MoE 60B on a Mac?
Yes — Mac Studio M5 Max (36 GB) and 65 other Macs can run Mixtral 34Bx2 MoE 60B, the smallest at Q2_K. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 37.3 GB of usable unified memory (RAM minus macOS overhead), from Mac Studio M5 Max (48 GB).
Can I run Mixtral 34Bx2 MoE 60B locally?
Yes — Mixtral 34Bx2 MoE 60B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Mixtral 34Bx2 MoE 60B?
At Q4_K_M, Mixtral 34Bx2 MoE 60B can reach ~129 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 37.3 × 0.65 = ~139 tok/s
Estimated speed at Q4_K_M (37.3 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Mixtral 34Bx2 MoE 60B?
At Q4_K_M, the download is about 36.49 GB. The full-precision BF16 version is 121.63 GB. The smallest option (Q2_K) is 25.85 GB.
Which GPUs can run Mixtral 34Bx2 MoE 60B?
No single consumer GPU has enough VRAM to run Mixtral 34Bx2 MoE 60B at Q4_K_M (37.3 GB). Multi-GPU or professional hardware is required.
Which devices can run Mixtral 34Bx2 MoE 60B?
67 devices with unified memory can run Mixtral 34Bx2 MoE 60B at Q4_K_M (37.3 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.