Can Mixtral 8x22B run on NVIDIA B200 180GB?

YES — Runs Great

B68Good

Estimated from fit model

Mixtral 8x22B needs ~108.3 GB VRAM. NVIDIA B200 180GB has 180.0 GB. With Q4_K_M quantization, expect ~162 tok/s.

Runtime: llama.cppCapacity: RoomyBandwidth: HighStack: StandardBottleneck: Balanced

Operating mode

Choose the run profile you care about

Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.

Current mode

Balanced

Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 108.3 GB, 162.4 tok/s, Runs well

108.3 GB required180.0 GB available

60% VRAM used

Fit status

Runs well

Decode

162.4 tok/s

TTFT

1192 ms

Safe context

66K

Memory

108.3 GB / 180.0 GB

Memory breakdown

Weights86.0 GB

KV Cache3.4 GB

Runtime0.9 GB

Headroom18.0 GB

See how fast it feels

See how fast it feelsMixtral 8x22B on NVIDIA B200 180GB
1st promptCold start — includes initialization
>Solve: If a train travels at 60 mph for 2.5 hours, how far does it go?
Let me work through this step by step.

Given information:
  - Speed = 60 mph (miles per hour)
  - Time  = 2.5 hours

Formula: Distance = Speed x Time

Calculation:
  Distance = 60 mph x 2.5 h
           = 150 miles

The train travels 150 miles.

Double-check: at 60 mph the train covers 60 miles each hour. In half an hour it covers 30 miles. So in 2.5 hours: 60 + 60 + 30 = 150 miles. Confirmed.
2nd promptWarm — model ready, no init overhead
>What if the train stops for 20 minutes halfway?
The stop doesn't change the total distance covered while moving, but it affects total trip time.

Breakdown:
  - First half: 1.25 hours at 60 mph = 75 miles
  - Stop: 20 minutes (no distance)
  - Second half: 1.25 hours at 60 mph = 75 miles

Total distance: 150 miles (same)
Total time: 2.5 h + 0.33 h = 2.83 hours
Average speed: 150 / 2.83 = 53 mph

The distance stays the same but average speed drops to 53 mph because of the stop.
Estimated: 162.4 tok/s decode · 1.2s TTFT (warm) · 406 tok/s prefill

What limits this setup

This setup is broadly balanced for this model.

No major red flags

This recommendation has enough memory headroom and acceptable estimated speed for the selected workload.

Best improvement path

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	B	Runs well	162.4 tok/s	650 ms	66K
Coding	B	Runs well	162.4 tok/s	1192 ms	66K
Agentic Coding	B	Runs well	162.4 tok/s	1734 ms	66K
Reasoning	B	Runs well	162.4 tok/s	1409 ms	66K
RAG	B	Runs well	162.4 tok/s	2168 ms	66K

Inference speed

Mixtral 8x22B inference speed — tokens per second by GPU & Mac

Estimated decode speed (tokens/sec) for Mixtral 8x22B at Q4_K_M across popular GPUs and Apple Silicon, including multi-GPU rigs, using the fastest local runtime per device. Fastest is Mac Studio M3 Ultra 256GB at ~14 tok/s. Speed is memory-bandwidth bound, so cards that fit the whole model in VRAM run far faster than ones that offload to system RAM.

GPU / Mac	Memory	Quant	Speed (tok/s)	Fits?
Mac Studio M3 Ultra 256GB	256 GB	Q4_K_M	13.5	Fits
Mac Studio M2 Ultra 128GB	128 GB	Q4_K_M	9.2	Heavy offload
Mac Studio M1 Ultra 128GB	128 GB	Q4_K_M	8.7	Heavy offload
MacBook Pro M4 Max 128GB	128 GB	Q4_K_M	6.8	Heavy offload
2× RX 7900 XTX 24GB	48 GB	Q4_K_M	5.0	Too big
RTX 5090 32GB	32 GB	Q4_K_M	3.9	Too big
MacBook Pro M4 Max 64GB	64 GB	Q4_K_M	3.7	Too big
2× RTX 4090 24GB	48 GB	Q4_K_M	3.4	Too big
2× RTX 3090 24GB	48 GB	Q4_K_M	2.9	Too big
MacBook Pro M3 Max 64GB	64 GB	Q4_K_M	2.6	Too big
4× RTX 3060 12GB	48 GB	Q4_K_M	2.6	Too big
RTX 4090 24GB	24 GB	Q4_K_M	2.5	Too big
RX 7900 XTX 24GB	24 GB	Q4_K_M	2.5	Too big
MacBook Pro M1 Max 64GB	64 GB	Q4_K_M	2.4	Too big
MacBook Pro M4 Pro 48GB	48 GB	Q4_K_M	2.3	Too big
RTX 3090 24GB	24 GB	Q4_K_M	2.1	Too big
RTX 4080 Super 16GB	16 GB	Q4_K_M	2.0	Too big
RTX 4070 12GB	12 GB	Q4_K_M	2.0	Too big
RTX 3060 12GB	12 GB	Q4_K_M	2.0	Too big
RTX 4060 8GB	8 GB	Q4_K_M	2.0	Too big

Estimates for single-stream decoding at Q4_K_M; real tokens/sec varies with prompt length, context, batch size, and runtime build. Prompt processing (prefill) is faster than the decode figures shown here.

Quantization options

How Mixtral 8x22B (141B params) fits at each quantization level on NVIDIA B200 180GB (180.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	55.0 GB	Low	B56
Q3_K_S	3	69.1 GB	Low	B58
NVFP4	4	79.0 GB	Medium	B59
Q4_K_M	4	86.0 GB	Medium	B60
Q5_K_M	5	101.5 GB	High	B61
Q6_K	6	115.6 GB	High	B61
Q8_0Best for your GPU	8	150.9 GB	Very High	B61
F16	16	289.0 GB	Maximum	F0

Get started

Copy-paste commands to run Mixtral 8x22B on your machine.

Run

ollama run mixtral:8x22b

Frequently asked questions

Can NVIDIA B200 180GB run Mixtral 8x22B?

Yes, NVIDIA B200 180GB can run Mixtral 8x22B with a B grade (Runs well). Expected decode speed: 162.4 tok/s.

How much VRAM does Mixtral 8x22B need?

Mixtral 8x22B (141B parameters) requires approximately 108.3 GB of memory with Q4_K_M quantization.

What is the best quantization for Mixtral 8x22B?

The recommended quantization for Mixtral 8x22B is Q4_K_M, which balances quality and memory efficiency.

What speed will Mixtral 8x22B run at on NVIDIA B200 180GB?

On NVIDIA B200 180GB, Mixtral 8x22B achieves approximately 162.4 tokens per second decode speed with a time-to-first-token of 1192ms using Q4_K_M quantization.

Can NVIDIA B200 180GB run Mixtral 8x22B for coding?

For coding workloads, Mixtral 8x22B on NVIDIA B200 180GB receives a B grade with 162.4 tok/s and 66K context.

What context window can Mixtral 8x22B use on NVIDIA B200 180GB?

On NVIDIA B200 180GB, Mixtral 8x22B can safely use up to 66K tokens of context. The model's official context limit is 66K, but available memory constrains the safe maximum.

See all results for NVIDIA B200 180GB See all hardware for Mixtral 8x22B

Embed this result▼

Paste this snippet into any page to show a live fit card.

<iframe src="https://willitrunai.com/embed/mixtral-8x22b-on-b200-180gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: