Thealth Mixtral 8x7B — Hardware Requirements & GPU Compatibility
Chat
Thealth Mixtral 8x7B is a 8x7B-parameter open language model from TachyHealth in the Mixtral family. At BF16 it needs about 123.20 GB of VRAM — see which GPUs and Macs can run it below.
Thealth Mixtral 8x7B runs on a 128 GB+ Mac or a datacenter GPU
Best local pick → Mac Pro M2 Ultra (128 GB) · BF16 · 127.6 GB · ~14 tok/s
Macs from 128 GB (BF16 · ~14 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Thealth Mixtral 8x7B
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 TachyHealth/Thealth_Mixtral-8x7B --max-model-len 16384 --tensor-parallel-size 8vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on B200 (192 GB) (64 GB headroom) · BF16
- Generation speed
- ~106 tok/s
- generation speed
- Cost per 1M output tokens
- $15.70
- per 1M output tokens
How Much VRAM Does Thealth Mixtral 8x7B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | Runs on | |
|---|---|---|---|---|
| 123.2 GB | 132.7 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Runs on Instinct MI250X 128 GB |
Selected BF16
Brain floating point 16127.6GBat 16K context
- Weights
- 122.3 GB
- KV cache
- 5.1 GB
- Overhead
- 0.3 GB
Server-class: ~127.6 GB at 16K context · one Instinct MI250X 128 GB
Or .
Which Hardware Can Run Thealth Mixtral 8x7B?
BF16 · 123.2 GBThealth Mixtral 8x7B (BF16) requires 123.2 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 161+ GB is recommended. No consumer GPU has enough memory. 15 devices with unified memory can run Thealth Mixtral 8x7B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Pro M2 Ultra (128 GB).
128–144 GB
fits · 0.4 GB free · up to 16K ctx
4 GPUs · 2 Mac classes
GH200 Grace Hopper Superchip~73Simulate output on GH200 Grace Hopper SuperchipH200 NVL~72Simulate output on H200 NVLH200 SXM~72Simulate output on H200 SXMM2 Ultra*128 GB~14Simulate output on Mac Pro M2 Ultra (128 GB)M5 Max128 GB~11Simulate output on Mac Studio M5 Max (128 GB)192+ GB
fits · 64.4 GB free · up to 32K ctx
6 GPUs · 1 Mac class · 2 device families
B200~106Simulate output on B200B300~106Simulate output on B300Instinct MI350X~59Simulate output on Instinct MI350XM3 Ultra*256–512 GB~14Simulate output on Mac Studio (M3 Ultra, 256GB)8x H100 SXM5~214Simulate output on NVIDIA DGX H1008x A100 SXM4~167Simulate output on NVIDIA DGX A100 640GB
By VRAM tier · typical GPUs and Macs at BF16
Consumer GPUs (up to 32 GB) and Macs up to 96 GB: too small for this model (needs ≥ 127.6 GB)8 GB tier →12 GB tier →16 GB tier →24 GB tier →Macs →
Alternatives to Thealth Mixtral 8x7B
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
- TachyHealth
- Family
- Mixtral
- Parameters
- 8x7B
- Active parameters
- 16.1B per token
- Release Date
- 2023-12-19
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Thealth Mixtral 8x7B need?
Thealth Mixtral 8x7B requires 123.2 GB of VRAM at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 8x7B × 16 bits ÷ 8 = 112 GB
KV Cache + Overhead ≈ 11.2 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.
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Thealth Mixtral 8x7B?
No — Thealth Mixtral 8x7B requires at least 123.2 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
Can I run Thealth Mixtral 8x7B on a Mac?
Yes — Mac Pro M2 Ultra (128 GB) and 14 other Macs can run Thealth Mixtral 8x7B. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 123.2 GB of usable unified memory (RAM minus macOS overhead).
Can I run Thealth Mixtral 8x7B locally?
Yes — Thealth Mixtral 8x7B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Thealth Mixtral 8x7B?
At BF16, Thealth Mixtral 8x7B can reach ~59 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 35.4 GB active ÷ (8000 × 0.65) = 6.81 ms, plus 48 layers × 0.055 ms = 2.64 ms, so 1000 ÷ 9.45 ms = ~106 tok/s
Estimated speed at BF16 (123.2 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Thealth Mixtral 8x7B?
At BF16, the download is about 112.00 GB.
Which GPUs can run Thealth Mixtral 8x7B?
No single consumer GPU has enough VRAM to run Thealth Mixtral 8x7B at BF16 (123.2 GB). Multi-GPU or professional hardware is required.
Which devices can run Thealth Mixtral 8x7B?
15 devices with unified memory can run Thealth Mixtral 8x7B at BF16 (123.2 GB), including Mac Pro M2 Ultra (128 GB), Mac Pro M2 Ultra (192 GB), Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.