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

Phi-4

  • Microsoft
  • Release · December 12, 2024
  • Open weights
  • Non-reasoning

Intelligence

Intelligence Index

#235 of 266
  1. 6.16.1Qwen3.5 0.8B (Reasoning)
  2. 66Ministral 3 14B
  3. 66Qwen3 Omni 30B A3B Instruct
  4. 66Granite 4.0 H Small
  5. 5.95.9Nova Micro
  6. 5.95.9Phi-4
  7. 5.95.9LFM2 24B A2B
  8. 5.85.8NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)
  9. 5.85.8Phi-4 Multimodal Instruct
  10. 5.75.7MiniCPM-V 4.6 1.3B
Data from Artificial Analysis

Evaluations

  • AA-Omniscience Accuracy14.1%#187
  • Humanity's Last Exam3.8%#248
  • AA-LCR0%#229
  • CritPt0%#143
  • GPQA Diamond57.5%#149
  • IFBench23.5%#143
  • τ²-Bench Telecom0%#143
  • Terminal-Bench Hard3.8%#90
Phi-4Median
Data from Artificial Analysis

Speed and price

Speed

tokens per second · #157 of 162
  1. 4545Gemma 4 E4B (Non-reasoning)
  2. 4343Hermes 4 - Llama-3.1 405B (Non-reasoning)
  3. 4343Hermes 4 - Llama-3.1 405B (Reasoning)
  4. 4141Qwen3.8 2.4T A95B
  5. 4040Qwen3.8 Max (0902)
  6. 3939Phi-4
  7. 3535Gemma 4 31B (Reasoning)
  8. 3333Command A
  9. 2626Llama 3.2 Instruct 11B (Vision)
  10. 1919Qwen3.5 4B (Reasoning)
Data from Artificial Analysis

Latency

seconds to first token · #98 of 162
  1. 2.32.3Hermes 4 - Llama-3.1 405B (Reasoning)
  2. 2.32.3GPT-6 Astra (Low)
  3. 2.42.4Mistral Medium 3.5
  4. 2.42.4GLM-5.3 (Max)
  5. 2.52.5GPT-6 Luna (Low)
  6. 2.52.5Phi-4
  7. 2.62.6LongCat 2.0
  8. 2.62.6Ling 3.0 Tiny
  9. 2.82.8Hy3
  10. 2.82.8GPT-6 Luna (Medium)
Data from Artificial Analysis

Price

$ per 1M tokens (3:1) · #49 of 169
  1. $0.20.2GPT-6 Luna (Medium)
  2. $0.20.2Claude Haiku 5.5 (Low)
  3. $0.20.2GPT-6 Luna (Low)
  4. $0.20.2GPT-6 Luna (Non-reasoning)
  5. $0.20.2Ministral 3 14B
  6. $0.220.22Phi-4
  7. $0.220.22Gemma 4 31B (Non-reasoning)
  8. $0.230.23Qwen3.8-Flash-Next
  9. $0.240.24GLM-5.3-Flash
  10. $0.240.24Hy3
Data from Artificial Analysis

Nearby models

  • Qwen3.5 0.8B (Reasoning)Alibaba6.1
  • Ministral 3 14BMistral6
  • Qwen3 Omni 30B A3B InstructAlibaba6
  • Granite 4.0 H SmallIBM6
  • Nova MicroAmazon5.9
  • LFM2 24B A2BLiquid AI5.9
  • NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)NVIDIA5.8
  • Phi-4 Multimodal InstructMicrosoft5.8
Data from Artificial Analysis

Same lab

  • Phi-4 Mini InstructMicrosoft6.3
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend