LLM Reference

Qwen2 Models by Alibaba

AlibabaApache 2.0Open sourceHighlight
8 models2024Up to 128k ctxFrom $0.05/1M input

Details

ResearcherAlibaba
LicenseApache 2.0(OSI)
Commercial useCommercial use allowed
Models8
Released2024
Max context128k

Capabilities

Structured Outputs5 of 8 models

Links

WebsiteHuggingFace

About

Qwen2 is a family of 8 AI models by Alibaba, released in 2024.

Current Variants

Use-when guidance is derived from seed capabilities, context, release, and replacement fields.

8 in view
Together AI Qwen2-7B-InstructCurrent

Use when the workload needs 33k context, 7B parameters, and structured outputs.

2024-0633k context7B parametersstructured outputs
Together AI Qwen2-72B-InstructCurrent

Use when the workload needs 33k context, 72B parameters, and structured outputs.

2024-0633k context72B parametersstructured outputs
Qwen2-7B-InstructCurrent

Use when the workload needs 128k context and 7B parameters.

2024-06128k context7B parameters
Qwen2-72BCurrent

Use when the workload needs 128k context, 72.7B parameters, and structured outputs.

2024-06128k context72.7B parametersstructured outputs
Qwen2-57B-A14BCurrent

Use when the workload needs 57.4B parameters and structured outputs.

2024-0657.4B parametersstructured outputs
Qwen2-7BCurrent

Use when the workload needs 128k context, 7.1B parameters, and structured outputs.

2024-06128k context7.1B parametersstructured outputs
Qwen2-1.5BCurrent

Use when the workload needs 1.5B parameters.

2024-061.5B parameters
Qwen2-0.5BCurrent

Use when the workload needs 490M parameters.

2024-06490M parameters

Release Timeline

1 release group
2024-06
8 current
Qwen2-0.5B
490M parameters
Current
Qwen2-1.5B
1.5B parameters
Current
Qwen2-57B-A14B
57.4B parametersstructured outputs
Current
Qwen2-72B
128k context72.7B parametersstructured outputs
Current
Qwen2-7B
128k context7.1B parametersstructured outputs
Current
Qwen2-7B-Instruct
128k context7B parameters
Current
Together AI Qwen2-72B-Instruct
33k context72B parametersstructured outputs
Current
Together AI Qwen2-7B-Instruct
33k context7B parametersstructured outputs
Current

Specifications(8 models)

Qwen2 model specifications comparison
ModelReleasedContextParametersStructured Outputs
Together AI Qwen2-7B-Instruct2024-0633k7BYes
Together AI Qwen2-72B-Instruct2024-0633k72BYes
Qwen2-7B-Instruct2024-06128k7BNo
Qwen2-72B2024-06128k72.71BYes
Qwen2-57B-A14B2024-06—57.41BYes
Qwen2-7B2024-06128k7.07BYes
Qwen2-1.5B2024-06—1.54BNo
Qwen2-0.5B2024-06—490MNo

Pricing

Qwen2 model pricing by provider
ModelProviderInput / 1MOutput / 1MType
Qwen2-7BDeepInfra$0.05$0.15Serverless
Qwen2-1.5BMicrosoft Foundry$0.07$0.07Provisioned
Qwen2-7BOctoAI API (Deprecated)$0.15$0.15Serverless
Qwen2-7BMicrosoft Foundry$0.15$0.15Provisioned
Together AI Qwen2-7B-InstructTogether AI$0.15$0.15Serverless
Qwen2-57B-A14BDeepInfra$0.16$0.16Serverless
Qwen2-7BFireworks AI$0.2$0.2Serverless
Qwen2-72BDeepInfra$0.45$0.65Serverless
Together AI Qwen2-72B-InstructTogether AI$0.7$0.7Serverless
Qwen2-72BFireworks AI$0.9$0.9Serverless
Qwen2-72BTogether AI$0.9$0.9Serverless
Qwen2-72BMicrosoft Foundry$1$2Provisioned

Frequently Asked Questions

What is Qwen2 used for?
Qwen2 is used for structured outputs, coding, and math-heavy prompts. The family description and listed model capabilities point to those workloads as the best fit.
How does Qwen2 compare to Tongyi DeepResearch?
Qwen2 by Alibaba is strongest where you need structured outputs, while Tongyi DeepResearch by Alibaba is the closest related family to check for adjacent model selection. Qwen2 has 8 listed variants and reaches up to 128k context, while Tongyi DeepResearch reaches up to 131k context, so compare the specs and pricing tables before choosing a production model.
Which Qwen2 model should I use?
For the lowest listed input price, start with Qwen2-7B through DeepInfra at $0.05/1M input tokens. For the most capable/latest local choice, evaluate Qwen2-72B with 128k context and structured outputs.