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

Gemini 3.8 Flash (High)

  • Google
  • Release · September 2, 2026
  • Proprietary
  • Reasoning

Intelligence

Intelligence Index

#43 of 266
  1. 42.142GPT-5.6 Terra (Max)
  2. 42.142GPT-6.1 Sol (Low)
  3. 41.842GLM-5.3-Flash
  4. 41.241Claude Haiku 5.5 (Xhigh)
  5. 41.141Ling 3.1 Flash
  6. 40.941Gemini 3.8 Flash (High)
  7. 40.841Claude Sonnet 5.5 (Medium)
  8. 39.940Qwen3.8 2.4T A95B
  9. 39.840Gemini 3.8 Flash (Medium)
  10. 39.840Qwen3.8-Flash-Next
Data from Artificial Analysis

Evaluations

  • AA-Briefcase1,202#24
  • AA-Omniscience Accuracy54.6%#23
  • AA-Omniscience Non-Hallucination44.8%#21
  • AutomationBench-AA59.9%#14
  • GDPval-AA1,436#19
  • Humanity's Last Exam47.8%#28
  • GDP.pdf21%#12
  • Terminal-Bench 4.019.7%#49
  • AA-LCR81.3%#34
  • SciCode56.6%#24
  • CritPt18.3%#41
  • Terminal-Bench-Science10%#13
  • FrontierCode36.4%#20
  • AA-AnalystAgent46.3%#10
  • Harvey LAB-AA0%#15
  • AA-3D966#12
  • ITBench-AA52.5%#2
  • MLCR-AA10.6%#22
  • MMMU-Pro85.6%#9
  • PersonalityBench: Expressiveness75.2#10
  • PersonalityBench: Organization67.8#13
  • PersonalityBench: Transparency18.3#24
  • PersonalityBench: Slop77.7#4
  • PersonalityBench: Informativeness82.7#7
  • PersonalityBench: Warmth53#14
  • PersonalityBench: Conviction82.7#5
  • PersonalityBench: Imagination35.6#21
  • GPQA Diamond95.3%#3
  • τ²-Bench Banking45%#11
Gemini 3.8 Flash (High)Median
Data from Artificial Analysis

Cost per task

Cost per task

$ per Intelligence Index task · #14 of 24
  1. $0.720.72GPT-6.1 Sol (Max)
  2. $0.840.84Muse Spark Open (Xhigh)
  3. $1.011.01Qwen3.8 27B (Xhigh)
  4. $1.031.03Step 5 Preview
  5. $1.131.13Mistral Large 4 Preview
  6. $1.241.24Gemini 3.8 Flash (High)
  7. $1.991.99Gemini 4 Argon (High)
  8. $2.012.01GLM-5.3 (Max)
  9. $2.112.11Kimi K3.1 Pro
  10. $3.263.26GPT-6 Astra (Max)
Data from Artificial Analysis

Time per task

seconds per task · #12 of 26
  1. 341341Muse Wonder (Max)
  2. 364364GPT-6 Luna (Max)
  3. 396396DeepSeek V4.1 Flash 1017 (Max)
  4. 412412GPT-6 Astra (Max)
  5. 428428Claude Haiku 5.5 (Max)
  6. 498498Gemini 3.8 Flash (High)
  7. 507507MiniMax-M3.1-Flash
  8. 578578Mistral Large 4 Preview
  9. 604604GLM-5.3 (Max)
  10. 628628GPT-6.1 Sol (Max)
Data from Artificial Analysis

Tokens per task

output tokens per task · #18 of 27
  1. 64.1K64KStep 5 Preview
  2. 64.3K64KMiMo-V2.6-Pro
  3. 66.8K67KQwen3.8 27B (Xhigh)
  4. 68.7K69KGLM-5.3-Flash
  5. 69.5K70KMistral Large 4 Preview
  6. 71K71KGemini 3.8 Flash (High)
  7. 71.1K71KGLM-5.3 (Max)
  8. 78.1K78KClaude Fable 5.1 (Max)
  9. 78.6K79KMiniMax-M3.1-Flash
  10. 80.6K81KGrok 4.7 (Xhigh)
Data from Artificial Analysis

Speed and price

Speed

tokens per second · #50 of 162
  1. 152152o3
  2. 151151Muse Spark Open (Xhigh)
  3. 147147Qwen3.6 35B A3B (Non-reasoning)
  4. 143143Qwen3.5 122B A10B (Non-reasoning)
  5. 140140GPT-6 Luna (Max)
  6. 137137Gemini 3.8 Flash (High)
  7. 136136Apertus 8B Instruct
  8. 134134Claude Sonnet 5.5 (Max)
  9. 133133Qwen3 Next 80B A3B Instruct
  10. 132132Gemini 3.1 Pro Preview
Data from Artificial Analysis

Latency

seconds to first token · #141 of 162
  1. 19.820GPT-6 Luna (Xhigh)
  2. 2020Claude Fable 5.1 (High)
  3. 20.621Gemini 3.1 Pro Preview
  4. 20.921Grok 4.7 (High)
  5. 2525Claude Haiku 5.5 (High)
  6. 27.327Gemini 3.8 Flash (High)
  7. 27.327Claude Sonnet 5.5 (Xhigh)
  8. 27.528Muse Wonder (Max)
  9. 30.330Claude Opus 5.5 (High)
  10. 36.136GPT-6.1 Sol (High)
Data from Artificial Analysis

Price

$ per 1M tokens (3:1) · #108 of 169
  1. $1.131.13Qwen3.8 27B (Non-reasoning)
  2. $1.31.3LongCat 2.0
  3. $1.351.35Qwen3.5 397B A17B (Non-reasoning)
  4. $1.351.35Qwen3.5 397B A17B (Reasoning)
  5. $1.431.43Step 5 Preview
  6. $1.51.5Gemini 3.8 Flash (High)
  7. $1.51.5Gemini 3.8 Flash (Medium)
  8. $1.51.5Gemini 3.8 Flash (Low)
  9. $1.51.5Qwen3.5 Omni Plus
  10. $1.51.5Hermes 4 - Llama-3.1 405B (Reasoning)
Data from Artificial Analysis

Nearby models

  • GPT-5.6 Terra (Max)OpenAI42.1
  • GPT-6.1 Sol (Low)OpenAI42.1
  • GLM-5.3-FlashZ AI41.8
  • Claude Haiku 5.5 (Xhigh)Anthropic41.2
  • Ling 3.1 FlashInclusionAI41.1
  • Claude Sonnet 5.5 (Medium)Anthropic40.8
  • Qwen3.8 2.4T A95BAlibaba39.9
  • Gemini 3.8 Flash (Medium)Google39.8
Data from Artificial Analysis

Same lab

  • Gemini 4 Argon (High)Google52.6
  • 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
  • Gemma 4 12B (Reasoning)Google14.2
  • Gemma 4 31B (Non-reasoning)Google13.9
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend