Compare LLM Models
Select 2-5 models to compare side by side.
Select models to compare
0 / 5 selected
Top downloads
Select 2-5 models
| Attribute |
🤖
sarvam-m
sarvamai
|
|---|---|
| Company | sarvamai |
| Release Date | 2025-05-23 |
| Parameters (B) | 24 |
| Architecture | Transformer |
| Context Window | 32768 |
| Input Modalities | text |
| Open Source | Yes |
| License | apache-2.0 |
| Score | 0 |
| VRAM (GB) | 48 |
| Compute | Single NVIDIA A100/H100 80GB GPU for BF16 inference; FP8 quantization reduces VRAM to ~24GB |
| Benchmark: ARC | 95 |
| Benchmark: GPQA-Diamond | 48 |
| Benchmark: GSM8K | 94 |
| Benchmark: HLE | 3.1 |
| Benchmark: HumanEval | 88 |
| Benchmark: IFEval | 87 |
| Benchmark: LiveCodeBench | 44 |
| Benchmark: MATH | 81 |
| Benchmark: MATH-500 | 84.7 |
| Benchmark: MBPP | 75 |
| Benchmark: MMLU | 87 |
| Benchmark: MMLU-Pro | 69.6 |
| Benchmark: SimpleQA | 5 |
| Pricing (Input) | — |
| Pricing (Output) | — |