Vascend
HorizonVascend IDEVascend CLICommunityBenchmarkContact
Sign inJoin Discord
HomeHorizonIDECLISign in
ChangelogIntelligenceCost per taskCoding Agent IndexCyber IndexDecision IndexBenchmarksSpeedLatencyPriceLeaderboard
Benchmark

Sarvam 30B (High)

  • Sarvam
  • Release · March 6, 2026
  • Open weights
  • Reasoning

Intelligence

Intelligence Index

#218 of 266
  1. 77Command A
  2. 6.96.9Qwen3.5 2B (Reasoning)
  3. 6.86.8NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)
  4. 6.86.8NVIDIA Nemotron Nano 9B V2 (Non-reasoning)
  5. 6.76.7Hermes 4 - Llama-3.1 70B (Non-reasoning)
  6. 6.66.6Sarvam 30B (High)
  7. 6.56.5Gemma 4 E2B (Non-reasoning)
  8. 6.56.5Llama 4 Scout
  9. 6.56.5Olmo 3.1 32B Instruct
  10. 6.46.4R1 1776
Data from Artificial Analysis

Evaluations

  • AA-Omniscience Accuracy12.6%#201
  • Humanity's Last Exam7.5%#172
  • AA-LCR0%#229
  • CritPt0.3%#117
  • GPQA Diamond63.3%#136
  • IFBench26.5%#137
  • τ²-Bench Telecom34.5%#80
  • Terminal-Bench Hard2.3%#100
Sarvam 30B (High)Median
Data from Artificial Analysis

Nearby models

  • Command ACohere7
  • Qwen3.5 2B (Reasoning)Alibaba6.9
  • NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)NVIDIA6.8
  • NVIDIA Nemotron Nano 9B V2 (Non-reasoning)NVIDIA6.8
  • Hermes 4 - Llama-3.1 70B (Non-reasoning)Nous Research6.7
  • Gemma 4 E2B (Non-reasoning)Google6.5
  • Llama 4 ScoutMeta6.5
  • Olmo 3.1 32B InstructAllen Institute for AI6.5
Data from Artificial Analysis

Same lab

  • Sarvam 105B (High)Sarvam8.8
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend