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

EXAONE 4.5 33B (Reasoning)

  • LG AI Research
  • Release · April 9, 2026
  • Open weights
  • Reasoning

Intelligence

Intelligence Index

#133 of 266
  1. 13.914Gemma 4 31B (Non-reasoning)
  2. 13.614Nova 2.0 Omni (Medium)
  3. 13.413Nova 2.0 Lite (High)
  4. 13.413Apriel-v1.6-15B-Thinker
  5. 13.313Qwen3.5 9B (Non-reasoning)
  6. 13.213EXAONE 4.5 33B (Reasoning)
  7. 13.113Qwen3.5 4B (Reasoning)
  8. 13.113Gemma 4 26B A4B (Non-reasoning)
  9. 13.113Command A+
  10. 12.913Nemotron 3.5 Lightning
Data from Artificial Analysis

Evaluations

  • AA-Omniscience Accuracy16.2%#162
  • Humanity's Last Exam12.9%#127
  • AA-LCR54.7%#138
  • CritPt0.3%#117
  • MMMU-Pro67.3%#68
  • GPQA Diamond79.4%#72
  • IFBench58%#45
  • τ²-Bench Telecom78.1%#44
  • Terminal-Bench Hard20.5%#36
EXAONE 4.5 33B (Reasoning)Median
Data from Artificial Analysis

Nearby models

  • Gemma 4 31B (Non-reasoning)Google13.9
  • Nova 2.0 Omni (Medium)Amazon13.6
  • Nova 2.0 Lite (High)Amazon13.4
  • Apriel-v1.6-15B-ThinkerServiceNow13.4
  • Qwen3.5 9B (Non-reasoning)Alibaba13.3
  • Qwen3.5 4B (Reasoning)Alibaba13.1
  • Gemma 4 26B A4B (Non-reasoning)Google13.1
  • Command A+Cohere13.1
Data from Artificial Analysis

Same lab

  • K-EXAONE 2.0 0803LG AI Research19.7
  • EXAONE 4.0 32B (Reasoning)LG AI Research8.2
  • EXAONE 4.0 32B (Non-reasoning)LG AI Research6.3
  • Exaone 4.0 1.2B (Reasoning)LG AI Research5.3
  • Exaone 4.0 1.2B (Non-reasoning)LG AI Research5.2
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend