text model · Sarvam · macOS

Can I run Sarvam-M 24B on Apple M5 (32GB)?

Compatibility verdict VRAM threshold engine

Yes, it runs GPU accelerated ~9 tok/s est.

Yes. Sarvam-M 24B runs on Apple M5 (32GB) at Q4_K_M (~16.3 GB of ~21 GB usable).

Needs ~16.3 GB Device usable ~21 GB

Runs at Q4_K_M using ~16.3 GB of ~21 GB usable.

That figure is at a 4k context and moves about ±15% as context length changes. Apple M5 (32GB) leaves ~4.7 GB of headroom.

Q4_K_M needed: ~16.3 GB
Usable on device: ~21 GB
Device memory: 32 GB
Best quant: Q4_K_M

Share on X Share on Reddit

Which quant fits

Quant ladder vs ~21 GB usable

Q2_K

~12.1 GB

Q3_K_M

~13.7 GB

Q4_K_M

~16.3 GB

Q5_K_M

~19.1 GB

Q6_K

~21.7 GB

Q8_0

~27.1 GB

FP16

~49.2 GB

The line marks Apple M5 (32GB)'s ~21 GB budget; rungs past it are too large.

Run it

Install commands macOS

Pick your tool. All 2 load the same Q4_K_M weights.

llama.cpp

$ llama-cli -hf lmstudio-community/sarvam-m-GGUF:Q4_K_M

LM Studio

$ lms get lmstudio-community/sarvam-m-GGUF

How to run it

On macOS use LM Studio (Polished GUI, ships MLX on Apple Silicon, one-click model downloads.).

Model Sarvam

Parameters: 24B
Q4_K_M size: 14.3 GB
Q8_0 size: 25.1 GB
Context: 32k

Full Sarvam-M 24B requirements →

Device macOS

Memory: 32 GB unified
Usable for weights: ~21 GB
Best runtime: MLX direct / Ollama (MLX backend)

Best models for Apple M5 (32GB) →

You could also run

Run Sarvam-M 24B on other hardware

Nvidia GeForce RTX 4090 (24GB)Nvidia GeForce RTX 3090 (24GB)AMD Radeon RX 7900 XTX (24GB)Nvidia GeForce RTX 5090 (32GB)32GB RAM Laptop (CPU/iGPU only)Apple M4 Pro (48GB)

FAQ

Can Apple M5 (32GB) run Sarvam-M 24B?

Yes. Sarvam-M 24B runs on Apple M5 (32GB) at Q4_K_M (~16.3 GB of ~21 GB usable).

How much memory does Sarvam-M 24B need?

Apple M5 (32GB) has room to spare. At Q4_K_M the weights are ~14.3 GB; with KV cache and runtime overhead, budget ~16.3 GB at a 4k context.

What is the best tool to run Sarvam-M 24B on macOS?

LM Studio for a simple setup; mlx-lm for the most speed. vLLM is NOT a Mac tool, it is a CUDA/Linux serving engine. Unified memory is not a fixed VRAM slice; ~70% is usable for weights.

Embed this

A live badge for your model card or README, updated as the data is.

Markdown

[![Sarvam-M 24B on Apple M5 (32GB)](https://localmodel.run/badge/sarvam-m-24b/apple-m5-32gb.svg)](https://localmodel.run/can-i-run/sarvam-m-24b/apple-m5-32gb)

Sources

Weights are measured from GGUF files; KV cache and overhead are computed, so totals can vary ~15% with context and runtime. Any tok/s is a bandwidth estimate. See methodology.