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

Solar Mini 4

  • Upstage
  • Release · September 22, 2026
  • Proprietary
  • Reasoning

Intelligence

Intelligence Index

#93 of 266
  1. 25.325K2 Horizon MoVA 36B A4B
  2. 25.225Qwen3.7 Plus
  3. 2525Inkling (Xhigh)
  4. 24.725Solar Open2 250B
  5. 24.625Ling-3.0-flash-VL
  6. 24.124Solar Mini 4
  7. 22.923Nemotron 3 Ultra 550B A55B (Reasoning)
  8. 22.623Ling-3.0-flash-Fin
  9. 22.222Gemini 3.5 Flash-Lite
  10. 21.822G9v3-39A5B
Data from Artificial Analysis

Evaluations

  • AA-Omniscience Accuracy18.4%#133
  • Humanity's Last Exam25.8%#90
  • Terminal-Bench 4.01%#77
  • AA-LCR83.3%#13
  • SciCode47.6%#78
  • CritPt1.4%#99
Solar Mini 4Median
Data from Artificial Analysis

Speed and price

Speed

tokens per second · #79 of 162
  1. 115115Nova 2.0 Pro Preview (Non-reasoning)
  2. 114114GPT-5.6 Terra (Medium)
  3. 114114Gemma 4 12B (Reasoning)
  4. 114114Agnes 2.5 Pro Beta
  5. 113113MiniMax-M3.1-Flash
  6. 112112Solar Mini 4
  7. 109109Mistral Large 4 Preview
  8. 108108Claude Sonnet 5.5 (Xhigh)
  9. 108108Gemma 4 26B A4B (Non-reasoning)
  10. 106106Claude Sonnet 5.5 (Medium)
Data from Artificial Analysis

Latency

seconds to first token · #75 of 162
  1. 1.81.8Qwen3.5 Omni Flash
  2. 1.81.8Command A
  3. 1.81.8Ling 3.1 Flash
  4. 1.81.8Gemma 4 31B (Non-reasoning)
  5. 1.81.8Qwen3 Omni 30B A3B Instruct
  6. 1.91.9Solar Mini 4
  7. 1.91.9Qwen3 Omni 30B A3B (Reasoning)
  8. 1.91.9Qwen3.5 9B (Reasoning)
  9. 22Qwen3.5 Omni Plus
  10. 22Qwen3.6 35B A3B (Non-reasoning)
Data from Artificial Analysis

Price

$ per 1M tokens (3:1) · #34 of 169
  1. $0.150.15Ministral 3 8B
  2. $0.150.15Qwen3.5 9B (Reasoning)
  3. $0.160.16Gemma 4 26B A4B (Reasoning)
  4. $0.170.17Quasar 438B (Max, Based on GLM-5.2)
  5. $0.180.18MiMo-V2.6-Flash
  6. $0.180.18Solar Mini 4
  7. $0.190.19Gemma 4 26B A4B (Non-reasoning)
  8. $0.190.19Qwen3.5 9B (Non-reasoning)
  9. $0.20.2Claude Haiku 5.5 (Max)
  10. $0.20.2Claude Haiku 5.5 (Xhigh)
Data from Artificial Analysis

Nearby models

  • K2 Horizon MoVA 36B A4BInstitute of Foundation Models25.3
  • Qwen3.7 PlusAlibaba25.2
  • Inkling (Xhigh)Thinking Machines25
  • Solar Open2 250BUpstage24.7
  • Ling-3.0-flash-VLInclusionAI24.6
  • Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA22.9
  • Ling-3.0-flash-FinInclusionAI22.6
  • Gemini 3.5 Flash-LiteGoogle22.2
Data from Artificial Analysis

Same lab

  • Solar Pro 4Upstage26.6
  • Solar Open 100B (Reasoning)Upstage10.4
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend