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

Jamba Reasoning 3B

  • AI21 Labs
  • Release · October 8, 2025
  • Open weights
  • Reasoning

Intelligence

Intelligence Index

#240 of 266
  1. 5.95.9Phi-4
  2. 5.95.9LFM2 24B A2B
  3. 5.85.8NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)
  4. 5.85.8Phi-4 Multimodal Instruct
  5. 5.75.7MiniCPM-V 4.6 1.3B
  6. 5.75.7Jamba Reasoning 3B
  7. 5.65.6Olmo 3 7B Think
  8. 5.65.6Molmo 7B-D
  9. 5.65.6Reka Flash 3
  10. 5.55.5Ministral 3 8B
Data from Artificial Analysis

Evaluations

  • AA-Omniscience Accuracy6.7%#243
  • Humanity's Last Exam3.8%#248
  • AA-LCR6.3%#223
  • CritPt0%#143
  • GPQA Diamond33.3%#194
  • IFBench52.5%#56
  • τ²-Bench Telecom15.8%#123
  • Terminal-Bench Hard0.8%#115
Jamba Reasoning 3BMedian
Data from Artificial Analysis

Nearby models

  • Phi-4Microsoft5.9
  • LFM2 24B A2BLiquid AI5.9
  • NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)NVIDIA5.8
  • Phi-4 Multimodal InstructMicrosoft5.8
  • MiniCPM-V 4.6 1.3BOpenBMB5.7
  • Olmo 3 7B ThinkAllen Institute for AI5.6
  • Molmo 7B-DAllen Institute for AI5.6
  • Reka Flash 3Reka AI5.6
Data from Artificial Analysis

Same lab

  • Jamba 1.7 LargeAI21 Labs6.1
  • Jamba 1.7 MiniAI21 Labs5.2
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend