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

Gemma 4 12B (Non-reasoning)

  • Google
  • Open weights
  • Non-reasoning

Intelligence

Intelligence Index

#173 of 266
  1. 1010gpt-oss-20b (Low)
  2. 9.910North Mini Code
  3. 9.910K2-V2 (High)
  4. 9.610Qwen3 Next 80B A3B Instruct
  5. 9.510DiffusionGemma 26B A4B
  6. 9.49.4Gemma 4 12B (Non-reasoning)
  7. 9.39.3Llama 4 Maverick
  8. 9.29.2Qwen3 Coder Next
  9. 9.29.2Motif-2-12.7B-Reasoning
  10. 9.29.2Nova Premier
Data from Artificial Analysis

Evaluations

  • AA-Omniscience Accuracy11.9%#204
  • Humanity's Last Exam6.4%#181
  • AA-LCR35%#167
  • CritPt0%#143
  • MMMU-Pro62%#87
  • GPQA Diamond66.1%#130
  • IFBench45.2%#70
  • τ²-Bench Telecom31.9%#81
  • Terminal-Bench Hard11.4%#53
Gemma 4 12B (Non-reasoning)Median
Data from Artificial Analysis

Speed and price

Speed

tokens per second · #71 of 162
  1. 120120GPT-6 Luna (High)
  2. 119119Ministral 3 8B
  3. 119119GPT-5.6 Terra (High)
  4. 118118GPT-5.6 Terra (Xhigh)
  5. 118118Llama 4 Scout
  6. 118118Gemma 4 12B (Non-reasoning)
  7. 116116GPT-6 Luna (Low)
  8. 116116GPT-5.6 Terra (Low)
  9. 115115Nova 2.0 Pro Preview (Non-reasoning)
  10. 114114GPT-5.6 Terra (Medium)
Data from Artificial Analysis

Latency

seconds to first token · #88 of 162
  1. 2.22.2Qwen3.5 397B A17B (Non-reasoning)
  2. 2.22.2Qwen3.5 397B A17B (Reasoning)
  3. 2.22.2K2 Horizon 375B A23B
  4. 2.22.2GPT-5.6 Terra (High)
  5. 2.22.2Gemma 4 12B (Reasoning)
  6. 2.32.3Gemma 4 12B (Non-reasoning)
  7. 2.32.3Qwen3.5 122B A10B (Reasoning)
  8. 2.32.3Hermes 4 - Llama-3.1 405B (Non-reasoning)
  9. 2.32.3Qwen3.7 Plus
  10. 2.32.3Qwen3.5 122B A10B (Non-reasoning)
Data from Artificial Analysis

Price

$ per 1M tokens (3:1) · #28 of 169
  1. $0.130.13gpt-oss-20b (Low)
  2. $0.130.13Phi-4 Mini Instruct
  3. $0.140.14Phi-4 Multimodal Instruct
  4. $0.150.15Agnes 2.5 Pro Beta
  5. $0.150.15Gemma 4 12B (Reasoning)
  6. $0.150.15Gemma 4 12B (Non-reasoning)
  7. $0.150.15Ministral 3 8B
  8. $0.150.15Qwen3.5 9B (Reasoning)
  9. $0.160.16Gemma 4 26B A4B (Reasoning)
  10. $0.170.17Quasar 438B (Max, Based on GLM-5.2)
Data from Artificial Analysis

Nearby models

  • gpt-oss-20b (Low)OpenAI10
  • North Mini CodeCohere9.9
  • K2-V2 (High)Institute of Foundation Models9.9
  • Qwen3 Next 80B A3B InstructAlibaba9.6
  • DiffusionGemma 26B A4BGoogle9.5
  • Llama 4 MaverickMeta9.3
  • Qwen3 Coder NextAlibaba9.2
  • Motif-2-12.7B-ReasoningMotif Technologies9.2
Data from Artificial Analysis

Same lab

  • Gemini 4 Argon (High)Google52.6
  • Gemini 3.8 Flash (High)Google40.9
  • Gemini 3.8 Flash (Medium)Google39.8
  • Gemini 3.8 Flash (Low)Google33.5
  • Gemini 3.1 Pro PreviewGoogle29.7
  • Gemini 3.5 Flash-LiteGoogle22.2
  • Gemma 4 26B A4B (Reasoning)Google16.7
  • Gemma 4 31B (Reasoning)Google14.7
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend