Sarvam 105B is Sarvam's language model with a 128K context window, starting at $0.106 / 1M input and $0.266 / 1M output. A 105-billion-parameter large language model from Sarvam AI, designed for multilingual tasks with a focus on Indian languages.
| Benchmarks | |
|---|---|
| GPQA | #236 |
| HLE | #235 |
| IFBench | #322 |
| Time to First Token | #473 |
| SciCode | #389 |
| LCR | #493 |
| TerminalBench Hard | #349 |
| TAU2 | #198 |
| Output TPS | #10 |
Capabilities
Input1/5
Text✓
Image·
Audio·
Video·
PDF·
Output1/5
Text✓
Image·
Audio·
Video·
Embedding·
Capabilities5/13
Reasoning✓
Adaptive Reasoning·
Function Calling✓
Parallel Function Calling✓
Structured Outputs✓
Native JSON Schema·
Web Search·
URL Context·
Computer Use·
Code Execution·
File Search·
Prompt Caching✓
Assistant Prefill·
Pricing by Provider
US Dollar ($)
Per 1M tokens
| Provider | Standard | ||
|---|---|---|---|
| Input $ / 1M | Output $ / 1M | Cache Read $ / 1M | |
| $0.106 | $0.266 | $0.00266 | |
Cost Calculator
US Dollar ($)
Preset:
Cheapest Instances to Run It
Cloud GPU instances that can host Sarvam 105B, ranked by cheapest on-demand price. The model needs about 252 GB of GPU memory at FP16 precision (estimated from its parameter count), so treat the fit as guidance rather than a guarantee.
All clouds
FP16 (full precision)
US Dollar ($)
Instance | Cloud | GPU | VRAM | Price | Cheapest region | |
|---|---|---|---|---|---|---|
| Standard_NP40s | 4× AMD Alveo U250 FPGA (64GB) | 256 GB | $6.60/hr | |||
| Standard_NC96ads_A100_v4 | 4× NVIDIA A100 | 320 GB | $14.69/hr | |||
| g7e.24xlarge | 4× RTX PRO Server 6000 | 384 GB | $16.57/hr | |||
Versions
| Version | Released | Context | Input / 1M | Output / 1M | Status |
|---|---|---|---|---|---|
| Sarvam 105B | — | 128K | $0.106 | $0.266 | Current |
| Sarvam 105B Conversations | — | 32K | $0.106 | $0.266 | Available |
| Sarvam 30B | — | — | — | — | Available |
| Sarvam M | — | 8K | — | — | Deprecated |
| Sarvam M Reasoning | — | — | — | — | Available |