본문 바로가기

벤치마크

Artificial Analysis의 지능·코딩·에이전트 지수와 Design Arena ELO를 출처별로 분리합니다. 실사용량 순위나 LastAI 점수에 합산하지 않습니다.

#1 Google: Gemini 4 Neon · 1500.0

#2 Google: Gemini 4 Neon (batch) · 1500.0

#3 Anthropic: Claude Fable 5.2 · 1486.0

#4 Anthropic: Claude Fable 5.2 (batch) · 1486.0

#5 Google: Gemini 3.8 Flash TTS · 1479.0

#6 Google: Gemini 4 Hydrogen · 1453.0

#7 Google: Gemini 4 Hydrogen (batch) · 1453.0

#8 Anthropic: Claude Opus 5.5 · 1449.0

#9 Anthropic: Claude Opus 5.5 (batch) · 1449.0

#10 MoonshotAI: Kimi K3.1 Pro · 1432.0

#11 MoonshotAI: Kimi K3.1 Pro (batch) · 1432.0

#12 Meta: Muse Wonder · 1420.0

#13 OpenAI: GPT-6.1 Sol · 1414.0

#14 OpenAI: GPT-6.1 Sol (batch) · 1414.0

#15 SpaceXSI: Grok 4.7 · 1413.0

#16 SpaceXSI: Grok 4.7 (batch) · 1413.0

#17 Google: Gemini 4 Carbon · 1412.0

#18 Google: Gemini 4 Carbon (batch) · 1412.0

#19 Sourceful: Riverflow V2.5 Pro · 1411.0

#20 Anthropic: Claude Fable 5.1 · 1394.0

Source: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings).

Source: OpenRouter · 2026-11-20T05:20:08.656Z · 2026-11-13 – 2026-11-19

마이