Apple192 GB VRAM800 GB/s$0/hr

Mistral Large 3 (256K) on Mac M2 Ultra (192GB unified): Does It Fit, How Fast, What It Costs

Mistral Large 3 (256K) at AWQ 4-bit needs 129.3 GB for its weights and 150.2 GB in total at 8K context with 8 concurrent users. That means it fits on a single Mac M2 Ultra (192GB unified). Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
150.2 GB needed / 192 GB available
Decode speed
11.72
tok/s per stream
Aggregate
95.07
tok/s @ 8 concurrent
Cost / 1M tokens
—
1× Mac M2 Ultra (192GB unified) · quote required
What changes as you add Mac M2 Ultra (192GB unified)s
Tensor-parallel degree vs VRAM, throughput, and cost. Cost scales linearly with GPU count; decode throughput scales near-linearly too, so cost per token stays roughly flat.
GPUsTotal VRAMVRAM neededDecode tok/sAggregate tok/sTTFTCost / 1M tokFits?
1× Mac192 GB150.2 GB11.7295.0786.28 s—✓
2× Mac384 GB150.2 GB19.92380.343.16 s—✓
4× Mac768 GB150.2 GB39.84152121.60 s—✓
8× Mac1536 GB150.2 GB79.68608510.81 s—✓
Concurrency vs context length
On 1× Mac M2 Ultra (192GB unified) at AWQ 4-bit. KV cache scales linearly with context, so this is the number of users your GPUs can actually serve.
ContextConcurrent usersKV per requestFits?
4K471.31 GB✓ fits
8K232.62 GB✓ fits
33K510.47 GB✓ fits
66K220.94 GB✓ fits
131K141.88 GB✓ fits
Other GPUs that run Mistral Large 3 (256K)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB8×17135$0.0648
L40S 48GB4×2065$0.3766
A100 80GB SXM42×1195$0.5578
H100 SXM5 80GB2×2002$0.6937
AMD MI300X 192GB1×931.6$0.7454
H200 SXM5 141GB2×2869$0.7747
Frequently asked questions

Can Mistral Large 3 (256K) run on Mac M2 Ultra (192GB unified)?

Yes. Mistral Large 3 (256K) at AWQ 4-bit needs 150.2 GB including the KV cache, and Mac M2 Ultra (192GB unified) has 192 GB — so it fits on a single GPU.

How fast is Mistral Large 3 (256K) on Mac M2 Ultra (192GB unified)?

11.72 tokens/sec per stream and 95.07 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 86.28 s, and inter-token latency is 85.3 ms.

What does Mistral Large 3 (256K) on Mac M2 Ultra (192GB unified) cost per million tokens?

— per 1M output tokens at 1× Mac M2 Ultra (192GB unified). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does Mistral Large 3 (256K) need on Mac M2 Ultra (192GB unified)?

129.3 GB for the weights at AWQ 4-bit, plus 20.9 GB of KV cache for 8 concurrent requests at 8K context — 150.2 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is Mac M2 Ultra (192GB unified) the right choice for Mistral Large 3 (256K)?

It works. Compare — per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 8× RTX 5090 32GB at $0.0648, and the fastest is 4× Cerebras CS-3 (WSE-3) at 47811 tok/s.

Adjust the workload yourself

Same formulas, your context length, concurrency, and quantisation.

Open in calculator