LLM Pricing
PricingLeaderboardsToolsProvidersReleasesGuides

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list
sa

Sarvam 30B

sarvam·sarvam/sarvam-30b·GA·Open weights·sarvam series
Efficient Indian-language reasoning model for chat, coding, and multilingual work

Specs & pricing

Input / output per 1M tokens
Official price
Price undisclosed
Lowest paid·FastRouterGateway
$0.02 / $0.10
Blended $0.04
Context
65,536
Output limit
65,536
Knowledge cutoff
—
Released / updated
2026-02-18 / 2026-02-18
Capabilities
✓ Reasoning✓ Tool use? Structured output✓ TemperatureAttachments
Modalities
Text

Quality & performanceArtificial Analysis · Intelligence Index v4.3 · rep. tier High

Intelligence6.6
Value165
Coding—
Agentic—
Output speed—
TTFT—
Value formulaIQ 6.6 ÷ min blended $0.040 = 165

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 2 providers1 with public prices

ProviderTierInputOutputCache readCache writeContextOutput limitStatus
FastRouterGateway$0.02$0.10——128,000 ⚠128,000
Sarvam AIOfficialFirst-party————65,53665,536hostTable.undisclosed

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Reasoning control

effort = nulleffort = loweffort = mediumeffort = high

Interleaved thinking (reasoning between tool calls) is declared by 1 of 2 providers.

1 / 2 providers expose no reasoning control (reasoning_options: []).

Your usage cost

1FastRouter$9.00
The cheapest paid channel is the only channel.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$0.04$02026-08-052026-08-132026-08-05 · Min blended $0.042026-08-06 · Min blended $0.042026-08-07 · Min blended $0.042026-08-08 · Min blended $0.042026-08-09 · Min blended $0.042026-08-10 · Min blended $0.042026-08-11 · Min blended $0.042026-08-12 · Min blended $0.042026-08-13 · Min blended $0.04

Artificial Analysis evaluations6 items

GPQA Diamond63.3%
Humanity's Last Exam7.5%
Terminal-Bench Hard2.3%
τ²-Bench Telecom34.5%
AA-LCR0.0%
IFBench26.5%

Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and harness, so rows are not comparable with one another. The Intelligence Index above draws on a different, newer set of evaluations.

Related models

saSarvam 105Bsame series—
Data partly from models.dev (MIT)