小規模オープンウェイトAIモデルの比較（4B～40B）

パラメーター数が4B～40BのオープンウェイトAIモデルです。

ウェイトをダウンロードできるモデルをオープンウェイト（一般にオープンソースとも呼ばれます）とみなします。独自のインフラストラクチャでセルフホストでき、ファインチューニングなどによるモデルのカスタマイズも可能です。

方法論などの詳細は、よくある質問をご覧ください。

AlibabaのロゴQwen3.8 27B (xhigh)とAlibabaのロゴQwen3.8 27B (medium)はパラメーター数4B～40Bと定義される小規模オープンウェイトモデルの中で知能が最も高く、AlibabaのロゴQwen3.8 27B (low)とAlibabaのロゴQwen3.6 27Bが続きます。

ハイライト

オープン性

Artificial Analysis Openness Index · Higher is better

知能

Artificial Analysis Intelligence Index · Higher is better

総パラメーター数

学習可能なパラメーター数（十億単位）

オープン性

Artificial Analysis Openness Index: Score

Openness Index assesses model openness on a 0 to 100 normalized scale (higher is more open)
Reasoning models are indicated by a lightbulb icon

知能

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index v4.1.1 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 v4.1.1 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 Evaluations

Intelligence evaluations measured independently by Artificial Analysis · Higher is better
See more
GDPval-AA v2

Agentic real-world work tasks, (Elo-500)/2000

𝜏³-BankingUpdated

Agentic tool use

Terminal-Bench v2.1

Agentic coding & terminal use

SciCode

Coding

Humanity's Last ExamUpdated

Reasoning & knowledge

GPQA Diamond

Scientific reasoning

CritPt

Physics reasoning

AA-Omniscience AccuracyUpdated

Knowledge

AA-Omniscience Non-Hallucination RateUpdated

1 - hallucination rate

AA-LCRUpdated

Long context reasoning

AA-Briefcase

Agentic knowledge work, Elo

AutomationBench-AA

Agentic SaaS workflows

No data available
Harvey LAB-AA

Legal agentic work, criterion pass rate

EnterpriseOps-Gym-AA

Agentic business operations

AA-AnalystAgentNew

Quantitative analysis on spreadsheets & documents

No data available
IFBench

Instruction following

APEX-Agents-AA

Long-horizon agentic tasks

No data available
ITBench-AA

Kubernetes incident root-cause analysis

MMMU-Pro

Visual reasoning

Reasoning models are indicated by a lightbulb icon

While model intelligence generally translates across use cases, specific evaluations may be more relevant for certain use cases.

Artificial Analysis Intelligence Index v4.1.1 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.

規模

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 v4.1.1 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 Index vs. Total Parameters

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 v4.1.1 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.

コンテキストウィンドウ

Context Window

Context window: tokens limit · Higher is better
Reasoning models are indicated by a lightbulb icon

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).

詳細

ウェイト
プロバイダーのベンチマーク
Qwen3.8 27B (xhigh)
AlibabaのロゴAlibaba
52
27B
256k
$0.4
53
🤗
Alibaba CloudSelf-hostedDeepInfra
+2
表示
Qwen3.8 27B (medium)
AlibabaのロゴAlibaba
44
27B
256k
$0.4
59
🤗
Alibaba Cloud
表示
Qwen3.8 27B (low)
AlibabaのロゴAlibaba
43
27B
256k
$0.4
64
🤗
Alibaba Cloud
表示
Qwen3.6 27B (Reasoning)
AlibabaのロゴAlibaba
38
27.8B
262k
$0.9
56
🤗
DeepInfraSiliconFlowGroq
+3
表示
Muse Glimmer (high)
MetaのロゴMeta
35
30B
131k
$0.2
109
🤗
Together AIDeepInfra
表示
Qwen3.8 27B (Non-reasoning)
AlibabaのロゴAlibaba
35
27B
256k
$0.4
64
🤗
Alibaba Cloud
表示
G9v3-39A5B
AI9StarsのロゴAI9Stars
34
39B
推論時に5Bが有効
131k
-
-
🤗
AI9Stars
表示
Qwen3.6 35B A3B (Reasoning)
AlibabaのロゴAlibaba
32
36B
推論時に3Bが有効
262k
$0.6
122
🤗
Self-hostedDeepInfraAlibaba Cloud
+5
表示
Qwen3.6 27B (Non-reasoning)
AlibabaのロゴAlibaba
31
27.8B
262k
$0.9
58
🤗
Multiverse ComputingGroqNovita
+2
表示
Gemma 4 31B (Reasoning)
GoogleのロゴGoogle
30
30.7B
256k
-
35
🤗
GMIGoogleLightning AI
+11
表示
Gemma 4 26B A4B (Reasoning)
GoogleのロゴGoogle
26
25.2B
推論時に3.8Bが有効
256k
$0.1
-
🤗
ParasailDeepInfraCloudflare
+5
表示
Qwen3.6 35B A3B (Non-reasoning)
AlibabaのロゴAlibaba
25
36B
推論時に3Bが有効
262k
$0.6
140
🤗
Self-hostedScalewayAlibaba Cloud
+4
表示