Sarvam AI/Dense

Sarvam 105B

Name: Sarvam 105B
Author: Sarvam AI

Large Indian multilingual model. Enterprise-grade quality across 10+ Indian languages.

chatmultilingualreasoning

105B

Parameters

32K

Context length

Benchmarks

Quantizations

Architecture

Dense

Released

2026-03-06

Layers

KV Heads

Head Dim

128

Family

other

Quantization Options

Quant	Bits	VRAM	Quality
IQ2_XXS	2.38	31.7 GB	low
IQ2_M	2.93	38.9 GB	low
Q2_K	3.16	42.0 GB	low
IQ3_XXS	3.25	43.1 GB	low
IQ3_XS	3.5	46.4 GB	low
Q3_K_S	3.64	48.3 GB	low
IQ3_M	3.76	49.8 GB	low
Q3_K_M	4	53.0 GB	low
Q3_K_L	4.3	56.9 GB	moderate
IQ4_XS	4.46	59.0 GB	moderate
Q4_K_S	4.67	61.8 GB	moderate
Q4_K_M	4.89	64.7 GB	good
Q5_K_S	5.57	73.6 GB	good
Q5_K_M	5.7	75.3 GB	good
Q6_K	6.56	86.6 GB	excellent
Q8_0	8.5	112.1 GB	lossless
FP16	16	210.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN SARVAM 105B NOW

Rent on RunPod →Or Vast.ai →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (13)

MATH98.6

AIME88.3

IFEval84.8

MMLU-PRO81.7

GPQA Diamond78.7

τ²-Bench46.8

SWE-bench45.0

IFBench34.4

SciCode26.4

AA Intelligence18.2

HLE10.1

AA Coding9.8

Terminal-Bench1.5

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run other:105b-q4_K_M

Tag may need adjustment — check ollama.com/library/other for available tags.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace GGUF Downloads Build Hardware