Skip to content

Head-to-head · Sarvam

Sarvam-1 2B vs Sarvam-30B

Sarvam-1 2B needs ~2.7 GB at Q4_K_M; Sarvam-30B needs ~21.7 GB. That ~19 GB gap decides which hardware runs each; they differ on 9 of the 14 devices below.

Spec Sarvam-1 2B Sarvam-30B
Parameters2B30B MoE
Memory at Q4_K_M~2.7 GB~21.7 GB
Context window8k64k
LMArenanot rankednot ranked
LicenseSarvam non-commercialApache-2.0

Which devices run each

A representative spread across the memory range. Tap a verdict for the full breakdown.

Size at each quantization

Quant Sarvam-1 2B Sarvam-30B
Q2_K 0.8 GB* 12.6 GB*
Q3_K_M 1 GB* 14.7 GB*
Q4_K_M 1.55 GB 19.6 GB
Q5_K_M 1.4 GB* 21.4 GB*
Q6_K 1.6 GB* 24.6 GB*
Q8_0 2.69 GB 31.9 GB*
FP16 5.05 GB 60 GB*

* derived from bits-per-weight; unstarred sizes are measured GGUF files.

Bottom line

The larger Sarvam-30B needs ~21.7 GB; Sarvam-1 2B runs on lighter hardware at ~2.7 GB with more headroom and faster responses. Check each against your exact device: what runs Sarvam-1 2B · what runs Sarvam-30B.

More Sarvam comparisons

FAQ

What is the difference in memory between Sarvam-1 2B and Sarvam-30B?

At Q4_K_M, Sarvam-1 2B needs about 2.7 GB and Sarvam-30B needs about 21.7 GB, a difference of ~19 GB.

Should I run Sarvam-1 2B or Sarvam-30B?

Run Sarvam-30B if your hardware has the ~21.7 GB it needs and you want maximum quality; run Sarvam-1 2B (~2.7 GB) for lighter hardware, faster responses, and more memory headroom.

Full breakdowns: Sarvam-1 2B · Sarvam-30B · all models × devices.

Sources

Memory figures are estimates at Q4_K_M, validated 2026-08-03. See methodology.