Vergleich von Modellen mit offenen Gewichten

Vergleich und Analyse von KI-Modellen mit offenen Gewichten anhand zentraler Leistungsmetriken wie Qualität, Leistung, Inferenzgeschwindigkeit, Kontextfenster, Parameteranzahl und Lizenzdetails.

Modelle gelten als Modelle mit offenen Gewichten (häufig auch als Open Source bezeichnet), wenn ihre Gewichte zum Download verfügbar sind. Dadurch können sie auf der eigenen Infrastruktur gehostet und beispielsweise durch Fine-Tuning angepasst werden.

Weitere Details zu unserer Methodik finden Sie in unseren FAQs.

Logo von KimiKimi K3 (max) und Logo von Z AIGLM-5.2 (max) weisen unter den Modellen mit offenen Gewichten die höchste Intelligenz auf, gefolgt von Logo von DeepSeekDeepSeek V4 Flash 0731 (max) und Logo von KimiKimi K3 (low).

Wichtigste Ergebnisse

Offenheit

Artificial Analysis Openness Index · Higher is better

Intelligenz

Artificial Analysis Intelligence Index · Higher is better

Parameter insgesamt

Trainierbare Parameter in Milliarden

Offenheit

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

Fortschritt bei Modellen mit offenen Gewichten

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).

Intelligenzentwicklung von Sprachmodellen mit offenen Gewichten nach Modelllabor

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.

Intelligenzentwicklung von Modellen mit offenen Gewichten nach Größe

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.

  • Tiny: Less than or equal to 4B parameters. These are usually the smallest models in terms of resource demand.
  • Small: Less than 40B parameters.
  • Medium: Between 40B-150B parameters.
  • Large: Over 150B parameters.

Intelligenz

Artificial Analysis Intelligence Index

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.

Intelligence Evaluations

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

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

𝜏³-Banking

Agentic tool use

Terminal-Bench v2.1

Agentic coding & terminal use

SciCode

Coding

Humanity's Last Exam

Reasoning & knowledge

GPQA Diamond

Scientific reasoning

CritPt

Physics reasoning

AA-Omniscience Accuracy

Knowledge

AA-Omniscience Non-Hallucination Rate

1 - hallucination rate

AA-LCR

Long context reasoning

AA-Briefcase

Agentic knowledge work, Elo

AutomationBench-AA

Agentic SaaS workflows

Harvey LAB-AA

Legal agentic work, criterion pass rate

EnterpriseOps-Gym-AANew

Agentic business operations

IFBench

Instruction following

APEX-Agents-AA

Long-horizon agentic tasks

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 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.

Größe

Intelligence Index nach Modellgröße

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.

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 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 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.

Kontextfenster

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).

Weitere Details

Gewichte
Anbieter-Benchmarks
Kimi K3 (max)
Logo von KimiKimi
57
2.8T
104B während der Inferenz aktiv
1M
$2.3
37
🤗
Together AIParasailMakora
+8
Anzeigen
GLM-5.2 (max)
Logo von Z AIZ AI
51
753B
40B während der Inferenz aktiv
1M
$0.9
160
🤗
FireworksDeepInfraSelf-hosted
+13
Anzeigen
DeepSeek V4 Flash 0731 (Reasoning, Max Effort)
Logo von DeepSeekDeepSeek
50
284B
13B während der Inferenz aktiv
1M
$0.1
110
🤗
Self-hostedSiliconFlowFireworks
+3
Anzeigen
MiniMax-M3
Logo von MiniMaxMiniMax
44
428B
23B während der Inferenz aktiv
1M
$0.2
73
🤗
NebiusSiliconFlowSelf-hosted
+7
Anzeigen
MiMo-V2.5-Pro
Logo von XiaomiXiaomi
42
1.0T
42B während der Inferenz aktiv
1M
$0.2
67
🤗
XiaomiDeepInfraGMI
+2
Anzeigen
Inkling (xhigh)
Logo von Thinking MachinesThinking Machines
41
975B
41B während der Inferenz aktiv
1M
$0.7
80
🤗
Thinking MachinesDeepInfraSelf-hostedTogether AI
Anzeigen
Nemotron 3 Ultra 550B A55B (Reasoning)
Logo von NVIDIANVIDIA
38
550B
55B während der Inferenz aktiv
262k
$0.5
145
Nicht verfügbar
DeepInfraTogether AIGMI
+6
Anzeigen
Mistral Medium 3.5
Logo von MistralMistral
30
128B
256k
$1.2
134
🤗
Self-hostedMistral
Anzeigen
Gemma 4 31B (Reasoning)
Logo von GoogleGoogle
29
30.7B
256k
-
35
🤗
SambaNovaParasailGMI
+10
Anzeigen
gpt-oss-120b (high)
Logo von OpenAIOpenAI
24
117B
5.1B während der Inferenz aktiv
131k
$0.2
185
🤗
DeepInfraGroqMicrosoft Azure
+18
Anzeigen
Command A+
Logo von CohereCohere
23
218B
25B während der Inferenz aktiv
192k
-
209
🤗
Cohere
Anzeigen