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

DeepSeek V4.1 Flash 1017 (Max)

  • DeepSeek
  • Release · October 17, 2026
  • Open weights
  • Reasoning

Intelligence

Intelligence Index

#30 of 266
  1. 46.446Grok 4.7 (Xhigh)
  2. 46.346MiMo-V2.6-Pro
  3. 46.346Grok 4.7 (High)
  4. 45.846GPT-6 Astra (Low)
  5. 45.445Qwen3.8 Max (0902)
  6. 45.145DeepSeek V4.1 Flash 1017 (Max)
  7. 44.845GLM-5.3 (Max)
  8. 43.944Ling 3.1 Flash 1019
  9. 43.744Step 5 Preview
  10. 43.443Claude Haiku 5.5 (Max)
Data from Artificial Analysis

Evaluations

  • AA-Briefcase1,570#10
  • AA-Omniscience Accuracy42.9%#49
  • AA-Omniscience Non-Hallucination33.3%#23
  • AutomationBench-AA66.6%#9
  • GDPval-AA1,746#7
  • Humanity's Last Exam43.3%#37
  • GDP.pdf17.2%#19
  • Terminal-Bench 4.033.8%#31
  • AA-LCR82.7%#24
  • SciCode55.9%#27
  • CritPt24%#31
  • Terminal-Bench-Science12.9%#11
  • FrontierCode43.7%#10
  • MLCR-AA26.7%#6
  • MMMU-Pro80.6%#24
  • PersonalityBench: Expressiveness44.2#18
  • PersonalityBench: Organization84.8#7
  • PersonalityBench: Transparency87.3#7
  • PersonalityBench: Slop41.6#14
  • PersonalityBench: Informativeness67#11
  • PersonalityBench: Warmth18.8#24
  • PersonalityBench: Conviction94.9#2
  • PersonalityBench: Imagination31.5#23
DeepSeek V4.1 Flash 1017 (Max)Median
Data from Artificial Analysis

Cost per task

Cost per task

$ per Intelligence Index task · #7 of 24
  1. $0.130.13MiMo-V2.6-Pro
  2. $0.190.19Nemotron 3.5 Super VL
  3. $0.210.21Claude Haiku 5.5 (Max)
  4. $0.220.22MiniMax-M3.1-Flash
  5. $0.230.23Beam 1.0
  6. $0.260.26DeepSeek V4.1 Flash 1017 (Max)
  7. $0.280.28GLM-5.3-Flash
  8. $0.720.72GPT-6.1 Sol (Max)
  9. $0.840.84Muse Spark Open (Xhigh)
  10. $1.011.01Qwen3.8 27B (Xhigh)
Data from Artificial Analysis

Time per task

seconds per task · #9 of 26
  1. 165165Inkling (Xhigh)
  2. 263263Step 5 Preview
  3. 284284Muse Spark Open (Xhigh)
  4. 341341Muse Wonder (Max)
  5. 364364GPT-6 Luna (Max)
  6. 396396DeepSeek V4.1 Flash 1017 (Max)
  7. 412412GPT-6 Astra (Max)
  8. 428428Claude Haiku 5.5 (Max)
  9. 498498Gemini 3.8 Flash (High)
  10. 507507MiniMax-M3.1-Flash
Data from Artificial Analysis

Tokens per task

output tokens per task · #23 of 27
  1. 71K71KGemini 3.8 Flash (High)
  2. 71.1K71KGLM-5.3 (Max)
  3. 78.1K78KClaude Fable 5.1 (Max)
  4. 78.6K79KMiniMax-M3.1-Flash
  5. 80.6K81KGrok 4.7 (Xhigh)
  6. 89.5K90KDeepSeek V4.1 Flash 1017 (Max)
  7. 107.7K108KQwen3.8 Max (0902)
  8. 119.2K119KClaude Opus 5.5 (Max)
  9. 162.2K162KClaude Haiku 5.5 (Max)
  10. 197.4K197KClaude Sonnet 5.5 (Max)
Data from Artificial Analysis

Speed and price

Speed

tokens per second · #37 of 162
  1. 177177DeepSeek V4 Flash Vision (Max)
  2. 173173Mistral Small 4 (Non-reasoning)
  3. 173173Claude Haiku 5.5 (Low)
  4. 173173gpt-oss-20b (High)
  5. 173173Nova 2.0 Lite (Low)
  6. 169169DeepSeek V4.1 Flash 1017 (Max)
  7. 168168NVIDIA Nemotron Nano 9B V2 (Non-reasoning)
  8. 166166Command A+
  9. 162162Mistral Medium 3.5
  10. 162162Muse Wonder (Max)
Data from Artificial Analysis

Latency

seconds to first token · #49 of 162
  1. 1.11.1NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)
  2. 1.11.1Muse Glimmer (High)
  3. 1.11.1Gemma 4 31B (Reasoning)
  4. 1.21.2Step 5 Preview
  5. 1.21.2Nemotron 3 Ultra 550B A55B (Reasoning)
  6. 1.21.2DeepSeek V4.1 Flash 1017 (Max)
  7. 1.31.3Quasar 438B (Max, Based on GLM-5.2)
  8. 1.31.3Trinity Large Thinking
  9. 1.31.3MiniMax-M3
  10. 1.31.3Llama 3.2 Instruct 11B (Vision)
Data from Artificial Analysis

Price

$ per 1M tokens (3:1) · #76 of 169
  1. $0.430.43Qwen3 Omni 30B A3B (Reasoning)
  2. $0.430.43Qwen3 Omni 30B A3B Instruct
  3. $0.450.45Ling 3.1 Flash 1019
  4. $0.450.45Ling 3.1 Flash
  5. $0.480.48Qwen3 Coder Next
  6. $0.530.53DeepSeek V4.1 Flash 1017 (Max)
  7. $0.530.53MiniMax-M3
  8. $0.530.53Solar Pro 4
  9. $0.530.53Inkling Small
  10. $0.530.53Muse Glimmer (High)
Data from Artificial Analysis

Nearby models

  • Grok 4.7 (Xhigh)SpaceXAI46.4
  • MiMo-V2.6-ProXiaomi46.3
  • Grok 4.7 (High)SpaceXAI46.3
  • GPT-6 Astra (Low)OpenAI45.8
  • Qwen3.8 Max (0902)Alibaba45.4
  • GLM-5.3 (Max)Z AI44.8
  • Ling 3.1 Flash 1019InclusionAI43.9
  • Step 5 PreviewStepFun43.7
Data from Artificial Analysis

Same lab

  • DeepSeek V4 Pro 0813 (Max)DeepSeek36
  • DeepSeek V4 Flash Vision (Max)DeepSeek34.8
  • DeepSeek V4 Pro 0813 (Non-reasoning)DeepSeek20.4
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend