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

Mercury 2.5

  • Inception
  • Release · September 8, 2026
  • Proprietary
  • Reasoning

Intelligence

Intelligence Index

#143 of 266
  1. 12.813Nova 2.0 Pro Preview (Low)
  2. 12.713Granite 4.2 30B
  3. 12.513Nova 2.0 Lite (Medium)
  4. 12.513Qwen3.5 Omni Flash
  5. 12.513MiniCPM5-2B
  6. 12.312Mercury 2.5
  7. 12.212JT-MINI
  8. 11.812Nova 2.0 Lite (Low)
  9. 11.812Magistral Medium 1.2
  10. 11.712HyperNova 60B 2605 (High, Based on gpt-oss-120b)
Data from Artificial Analysis

Evaluations

  • AA-Omniscience Accuracy22.7%#101
  • Humanity's Last Exam11.8%#132
  • Terminal-Bench 4.00%#96
  • AA-LCR71.7%#96
  • SciCode38.5%#113
  • CritPt0%#143
Mercury 2.5Median
Data from Artificial Analysis

Speed and price

Speed

tokens per second · #3 of 162
  1. 1,6191,619Celeris-1
  2. 879879K2 Horizon 375B A23B
  3. 714714Mercury 2.5
  4. 446446Nemotron 3.5 Super VL
  5. 414414Gemini 3.5 Flash-Lite
  6. 363363Beam 1.0
  7. 353353Trinity Large Thinking
  8. 295295Nemotron 3.5 Lightning
  9. 290290Nova Micro
  10. 277277NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)
Data from Artificial Analysis

Latency

seconds to first token · #111 of 162
  1. 2.92.9Qwen3.8 2.4T A95B
  2. 33GLM-5.3 (Low)
  3. 3.13.1Kimi K3.1 Pro
  4. 3.13.1MiMo-V2.6-Pro
  5. 3.53.5Apertus 8B Instruct
  6. 3.63.6Mercury 2.5
  7. 3.73.7Ring-2.6-1T
  8. 3.83.8MiMo-V2.6-Flash
  9. 3.93.9Qwen3.8 27B (Xhigh)
  10. 3.93.9Qwen3.8 27B (Low)
Data from Artificial Analysis

Price

$ per 1M tokens (3:1) · #68 of 169
  1. $0.310.31Nemotron 3.5 Super VL
  2. $0.330.33Celeris-1
  3. $0.350.35Llama 3.2 Instruct 11B (Vision)
  4. $0.350.35Qwen3 Next 80B A3B Instruct
  5. $0.350.35Reka Flash 3
  6. $0.380.38Mercury 2.5
  7. $0.390.39Trinity Large Thinking
  8. $0.420.42Llama 4 Maverick
  9. $0.430.43Qwen3 Omni 30B A3B (Reasoning)
  10. $0.430.43Qwen3 Omni 30B A3B Instruct
Data from Artificial Analysis

Nearby models

  • Nova 2.0 Pro Preview (Low)Amazon12.8
  • Granite 4.2 30BIBM12.7
  • Nova 2.0 Lite (Medium)Amazon12.5
  • Qwen3.5 Omni FlashAlibaba12.5
  • MiniCPM5-2BOpenBMB12.5
  • JT-MINIChina Mobile12.2
  • Nova 2.0 Lite (Low)Amazon11.8
  • Magistral Medium 1.2Mistral11.8
Data from Artificial Analysis
HorizonVascend IDEOlympusTalosCoursesCommunityBenchmark
DiscordGitHubXLet’s talkLegalPrivacyCookiesData useSecurity
© 2026 Vascend Technologies · VAT IT04911550616
ascend