Skip to content
Model Pareto

Sarvam 30B

Sarvam AIOpen weightsartificialanalysis.ai ↗
Compare with nearest rivalsCompare with…
Input
$0.026/M tok
Output
$0.11/M tok
Blended
$0.047/M tok · 3:1
Speed
~120tok/sest.
Context
66Ktokens

Long-tail import from Artificial Analysis (2026-09-24); AA variant: Sarvam 30B (high). Price = AA median host price.

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.

Hardware
1× H200 141GB
FP8 weights
Throughput
~9,509 tok/s
many requests batched
Price
$0.040–0.33 /M tok
busy → light use
On your own machine
RTX 5090 32GB (int4, 300+ tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • FP8 weights (30B params, 2.4B active per token) + 25% KV-cache headroom ≈ 38 GB
  • 1× H200 141GB at $3.59–$7.91/GPU-hour on-demand (2026-09-24)
  • ~9,509 output tok/s aggregate at batch 256 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~194,808 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Writing & chat

MMMLUEstimated
~last of 5 (estimated)below measured range (<81.3%)
  1. #1Gemini 3.1 Pro92.6%
  2. #2Gemma 4 31B88.4%
  3. #4gpt-oss-120b81.3%
  4. ~#5Sarvam 30BEstimated<81.3%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Estimated
~#45 of 54 (estimated)~1312 Elo
  1. #1Claude Opus 5.51501
  2. #2Claude Fable 5.1Few votes1496
  3. #44INTELLECT-3Few votes1327
  4. ~#45Sarvam 30BEstimated1312
  5. #45Llama 4 Maverick1311
No published score — estimated from related benchmarks
LMArena Text (Elo)Estimated
~#46 of 54 (estimated)~1336 Elo
  1. #1Claude Opus 5.51509
  2. #2Claude Fable 5.11501
  3. #45Llama 3.1 Nemotron Ultra 253B v1Few votes1348
  4. ~#46Sarvam 30BEstimated1336
  5. #46Olmo 3.1 32B Instruct1330
No published score — estimated from related benchmarks
EQ-Bench Creative Writing v3 (Elo)Estimated
~#23 of 26 (estimated)~1412 Elo
  1. #1GPT-6 Astra2173
  2. #2Claude Fable 5.12162
  3. #22Mistral Large 31412
  4. ~#23Sarvam 30BEstimated1412
  5. #23Gemma 4 31B1368
No published score — estimated from related benchmarks

Hard reasoning

AIME (latest)Estimated
~#14 of 25 (estimated)~52.8%
  1. #1Inkling97.1%
  2. #2Muse Glimmer94.7%
  3. #13Phi-4-reasoning-plus78.0%
  4. ~#14Sarvam 30BEstimated52.8%
  5. #14Mistral Large 338.0%
No published score — estimated from related benchmarks
AA Intelligence IndexEstimated
~#63 of 75 (estimated)~9.6
  1. #1Claude Opus 5.557.6
  2. #2Claude Sonnet 5.556.0
  3. #62North Mini Code9.9
  4. ~#63Sarvam 30BEstimated9.6
  5. #63Mistral Large 39.3
No published score — estimated from related benchmarks

Agents

OSWorld-VerifiedEstimated
~#11 of 12 (estimated)~65.0%
  1. #1Kimi K384.8%
  2. #2Qwen3.8 27B84.3%
  3. #10Muse Glimmer65.9%
  4. ~#11Sarvam 30BEstimated65.0%
  5. #11Claude Haiku 4.550.7%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#3 of 5 (estimated)~53.8%
  1. #1GLM-5.3 Flash59.1%
  2. #2Kimi K358.3%
  3. ~#3Sarvam 30BEstimated53.8%
  4. #3GPT-5.6 Terra50.2%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#17 of 18 (estimated)~34.6%
  1. #1Atria Dawn Preview92.5%
  2. #2GPT-6 Astra91.5%
  3. #16Nemotron 3 Ultra44.4%
  4. ~#17Sarvam 30BEstimated34.6%
  5. #17Mistral Small 421.3%
No published score — estimated from related benchmarks
GDPval-AA (Elo)Estimated
~#50 of 87 (estimated)~772 Elo
  1. #1Claude Opus 5.51846
  2. #2Claude Fable 5.11735
  3. #49Muse Glimmer774
  4. ~#50Sarvam 30BEstimated772
  5. #50Qwen3.7 Plus756
No published score — estimated from related benchmarks
τ-Bench Banking (AA)Estimated
~#71 of 89 (estimated)~6.4%
  1. #1Muse Spark 1.350.5%
  2. #2GLM-5.350.3%
  3. #70Ministral 3 14B6.6%
  4. ~#71Sarvam 30BEstimated6.4%
  5. #71North Mini Code6.4%
No published score — estimated from related benchmarks