What GPU do I need to run sarvamai/sarvam-30b?
32.2B parameters, published in F32. View on Hugging Face
Set up sarvam-30bsarvam-30b is published by sarvamai on Hugging Face, with 14,683 downloads and 222 likes to date. It's a SarvamMoEForCausalLM model built for text-generation, published natively in F32.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision, plus a fixed overhead for KV-cache, activations, and fragmentation. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| FP32 | 119.8 GB | 143.7 GB | V100 | 5 | $0.935/hr |
| FP8 (quantized) | 29.9 GB | 35.9 GB | RTX 4090 | 1 | $0.485/hr |
| cheaper alt. | RTX 5060 Ti | 3 | $0.330/hr | ||
| INT4 (quantized) | 15.0 GB | 18.0 GB | RTX A5000 | 1 | $0.176/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run sarvam-30b at its published (F32) precision: 5× V100, at $0.187/hr per GPU ($0.935/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
sarvam-30b: common questions
Can sarvam-30b run on a single GPU?
No. At FP32 it needs 143.7 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest capable card in the live feed is a 32.0 GB V100, and it takes 5 of them.
Can sarvam-30b run in 16-bit instead of FP32?
Yes. Its published weights are FP32, 119.8 GB, or 143.7 GB once inference overhead is added. Loading the same weights in 16 bits halves that to 59.9 GB, or 71.9 GB with overhead. How much accuracy the cast costs is model-specific and is not measured here.
How many GPUs do I need to run sarvam-30b?
5 at FP32. It needs 143.7 GB of VRAM and the cheapest capable live offer is a 32.0 GB V100, so 5 of them come to $0.935/hr in total.
What is the least VRAM sarvam-30b can run in?
18.0 GB, at INT4 (quantized), which fits a 24 GB card, against 143.7 GB at FP32. That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
Related reading: V100 pricing and specs, and The best GPUs for AI, ranked.
