Artificial Analysis AI 트렌드

현재 AI의 현황과 발전을 이끄는 주요 트렌드를 분석합니다. 모델의 지능, 효율성, 아키텍처, 추론 속도와 비용, 학습 트렌드를 다룹니다.

AI 현황(주요 내용) - 2025년 연말

AI 환경과 AI의 방향을 결정하는 주요 트렌드 분석

보고서 보기

AI 발전

AI의 지속적인 발전과 주요 AI 기업별 위치를 추적합니다.

Frontier Language Model Intelligence, Over Time

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

주요 기술 기업의 기간별 자본 지출

분기별 자본 지출(십억 USD)

Represents major investments by tech companies in infrastructure, including AI hardware (like GPUs and data centers). Capex is a strong indicator of a company's commitment to AI development, as training and running frontier models requires significant computing resources.

Note: Capex data is sourced from publicly available financial reports, news articles, and primarily from the SEC.

Intelligence Index와 출시일 비교

Artificial Analysis Intelligence Index · Release date
Most attractive region

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

AI 연구소별 주요 모델

각 AI 연구소가 달성한 최고 Artificial Analysis Intelligence Index
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

효율성

AI 효율성이 어떻게 발전하는지 분석합니다. 특정 지능 수준을 달성하는 비용과 그 지능을 이용할 수 있는 속도가 어떻게 변화하는지 살펴봅니다.

Intelligence Index 구간별 언어 모델 추론 가격 추이

100만 토큰당 가격(USD, 캐시·입력·출력 토큰 가격을 7:2:1로 혼합). 구간에는 Artificial Analysis Intelligence Index v4.1을 사용합니다.
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Intelligence Index 구간별 언어 모델 출력 속도 추이

초당 출력 토큰 수. 구간에는 Artificial Analysis Intelligence Index v4.1을 사용합니다.
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

국가별 분석

AI는 전 세계적인 현상입니다. AI 발전이 어디에서 일어나고 있는지, 각국의 주요 모델이 어떻게 비교되는지 살펴봅니다. AI 개발을 주도하는 두 중심지인 미국과 중국을 심층 분석합니다.

국가별 프런티어 언어 모델 지능 추이

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

오픈 웨이트: 국가별 프런티어 언어 모델 지능 추이

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Leading Models by Country

Artificial Analysis Intelligence Index · Leading models
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

오픈 소스 모델

오픈 웨이트 모델은 배포 유연성을 제공하며 특정 사용 사례에 맞게 미세 조정할 수 있습니다. 주요 오픈 웨이트 모델을 분석하고 독점 모델과 지능을 비교합니다.

Progress in Open Weights vs. Proprietary Intelligence

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Indicates whether the model weights are available. Models are labelled as 'Commercial Use Restricted' if the weights are available but commercial use is limited (typically requires obtaining a paid license).

Artificial Analysis Intelligence Index by Open Weights / Proprietary

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Indicates whether the model weights are available. Models are labelled as 'Commercial Use Restricted' if the weights are available but commercial use is limited (typically requires obtaining a paid license).

모델 아키텍처

AI 모델의 아키텍처는 성능과 효율성에 영향을 줍니다. 이 섹션에서는 Mixture of Experts(MoE) 아키텍처의 인기 상승과 같은 모델 아키텍처 트렌드와 모델 역량의 관계를 살펴봅니다.

모델 아키텍처별 Intelligence Index와 출시일 비교

Artificial Analysis Intelligence Index, 오픈 웨이트 모델만 해당
Most attractive region

A model where only a subset of parameters ("experts") are active per input. Routing mechanisms select a few experts per forward pass, reducing computation while allowing the model to scale to many more parameters overall.

A model where all parameters are active for every input. Every forward pass involves the full network, making it computationally intensive but straightforward to train and deploy.

Model Size: Total and Active Parameters

Comparison between total model parameters and parameters active during inference
Reasoning models are indicated by a lightbulb icon

The total number of trainable weights and biases in the model, expressed in billions. These parameters are learned during training and determine the model's ability to process and generate responses.

The number of parameters actually executed during each inference forward pass, expressed in billions. For Mixture of Experts (MoE) models, a routing mechanism selects a subset of experts per token, resulting in fewer active than total parameters. Dense models use all parameters, so active equals total.

Intelligence Index vs. Active Parameters

Artificial Analysis Intelligence Index · Active parameters at inference time
Most attractive quadrant
Pareto line
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

The number of parameters actually executed during each inference forward pass, expressed in billions. For Mixture of Experts (MoE) models, a routing mechanism selects a subset of experts per token, resulting in fewer active than total parameters. Dense models use all parameters, so active equals total.

Intelligence Index와 총 매개변수 비교

Artificial Analysis Intelligence Index · Size in parameters (billions)
Most attractive quadrant
Pareto line
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

The total number of trainable weights and biases in the model, expressed in billions. These parameters are learned during training and determine the model's ability to process and generate responses.

분기별 컨텍스트 길이(토큰) 중앙값

컨텍스트 길이 중앙값(천 토큰)

Larger context windows are relevant to RAG (Retrieval Augmented Generation) LLM workflows which typically involve reasoning and information retrieval of large amounts of data.

Maximum number of combined input & output tokens. Output tokens commonly have a significantly lower limit (varied by model).

학습 분석

AI 모델 학습 트렌드 분석에서는 학습 규모의 변화와 학습 규모 및 모델 지능의 관계를 모두 살펴봅니다.

모델별 학습 토큰

학습 토큰 수(조)
No data available
Reasoning models are indicated by a lightbulb icon

The number of tokens used to train the model, represented in trillions.

Intelligence Index와 학습 토큰 비교

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR · 학습 토큰 수(조)
Most attractive quadrant
No data available
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

The number of tokens used to train the model, represented in trillions.