Sarvam-105B
sarvam-105b
Flagship Indian-language reasoning model for enterprise multilingual applications
Specification
most-agreed values
Context
131K
Max output
131K
Released
2026-02-18
Knowledge cutoff
—
Retires
—
Open weights
yes
Input
text
Output
text
Price
US dollars per million tokens · most-agreed
Input
$0.1579
Output
$0.6316
Cache read
$0.0526
Available from 3 hosts
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| fastrouter · sarvam | $0.04 | $0.16 | $0.025 | — | 131K | 131K | — |
| nano-gpt | $0.054 | $0.2124 | $0.0336 | — | 131K | 4K | — |
| sarvam | $0.1579 | $0.6316 | $0.0526 | — | 131K | 131K | — |
Price history
append-only observations · a listing writes a row only when its price moves
| Date | Host | Field | $/M |
|---|---|---|---|
| 2026-09-02 | nano-gpt | Cache read | $0.028 |
| 2026-09-09 | nano-gpt | Cache read | $0.0336 |
| 2026-09-02 | nano-gpt | Input | $0.045 |
| 2026-09-09 | nano-gpt | Input | $0.054 |
| 2026-09-02 | nano-gpt | Output | $0.177 |
| 2026-09-09 | nano-gpt | Output | $0.2124 |