LLM Reference

Qwen3-Coder Models by Alibaba

AlibabaApache 2.0Open source
5 models2025–2026Up to 1m ctxFrom $0.07/1M input

Details

ResearcherAlibaba
LicenseApache 2.0OSI-approved
Commercial useCommercial use: permitted
Models5
Released2025–2026
Max context1m

Capabilities

Reasoning1 of 5 models
Function CallingAll models
Tool UseAll models
Structured OutputsAll models
Code Execution3 of 5 models

About

Qwen3-Coder is a family of 5 AI models by Alibaba, released between 2025 and 2026.

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

5 in view
Qwen3-Coder-NextCurrent

Use when the workload needs code, 256k context, and reasoning.

2026-02code256k contextreasoning
Qwen3-Coder-30B-A3B-InstructCurrent

Use when the workload needs code, 262k context, and tool use.

2025-12code262k contexttool use
Qwen3-Coder-PlusCurrent

Use when the workload needs code, 1m context, and tool use.

2025-09code1m contexttool use
Qwen3-Coder-FlashCurrent

Use when the workload needs code, 1m context, and tool use.

2025-07code1m contexttool use
Qwen3-Coder-480B-A35B-InstructCurrent

Use when the workload needs code, 262k context, and tool use.

2025-07code262k contexttool use

Release Timeline

4 release groups
2026-02
1 current
Qwen3-Coder-Next
code256k contextreasoning
Current
2025-12
1 current
Qwen3-Coder-30B-A3B-Instruct
code262k contexttool use
Current
2025-09
1 current
Qwen3-Coder-Plus
code1m contexttool use
Current
2025-07
2 current
Qwen3-Coder-480B-A35B-Instruct
code262k contexttool use
Current
Qwen3-Coder-Flash
code1m contexttool use
Current

Specifications(5 models)

Qwen3-Coder model specifications comparison
ModelReleasedContextParametersReasoningFn CallingTool UseStructured OutputsCode Exec
Qwen3-Coder-Next2026-02256k80B total, 3B activeYesYesYesYesYes
Qwen3-Coder-30B-A3B-Instruct2025-12262k30.5B total, 3.3B activeNoYesYesYesYes
Qwen3-Coder-Plus2025-091m—NoYesYesYesNo
Qwen3-Coder-Flash2025-071m—NoYesYesYesNo
Qwen3-Coder-480B-A35B-Instruct2025-07262k480B total, 35B activeNoYesYesYesYes

Pricing

Qwen3-Coder model pricing by provider
ModelProviderInput / 1MOutput / 1MType
Qwen3-Coder-30B-A3B-InstructNovita AI$0.07$0.27Serverless
Qwen3-Coder-NextOpenRouter$0.12$0.8Serverless
Qwen3-Coder-30B-A3B-InstructAWS Bedrock$0.15$0.62Serverless
Qwen3-Coder-30B-A3B-InstructVercel AI Gateway$0.15$0.6Serverless
Qwen3-Coder-NextNovita AI$0.2$1.5Serverless
Qwen3-Coder-480B-A35B-InstructGCP Vertex AI$0.22$1.8Serverless
Qwen3-Coder-FlashAlibaba Cloud PAI-EAS$0.3$1.5Serverless
Qwen3-Coder-480B-A35B-InstructNovita AI$0.38$1.55Serverless
Qwen3-Coder-NextAWS Bedrock$0.5$1.2Serverless
Qwen3-Coder-NextVercel AI Gateway$0.5$1.2Serverless
Qwen3-Coder-PlusAlibaba Cloud PAI-EAS$1$5Serverless
Qwen3-Coder-PlusVercel AI Gateway$1$5Serverless
Qwen3-Coder-480B-A35B-InstructVercel AI Gateway$1.5$7.5Serverless

Frequently Asked Questions

What is Qwen3-Coder used for?
Qwen3-Coder is used for code, reasoning, and agent workflows and tool use. The family description and listed model capabilities point to those workloads as the best fit.
How does Qwen3-Coder compare to Tongyi DeepResearch?
Qwen3-Coder by Alibaba is strongest where you need code, while Tongyi DeepResearch by Alibaba is the closest related family to check for adjacent model selection. Qwen3-Coder has 5 listed variants and reaches up to 1m context, while Tongyi DeepResearch reaches up to 131k context, so compare the specs and pricing tables before choosing a production model.
Which Qwen3-Coder model should I use?
For the lowest listed input price, start with Qwen3-Coder-30B-A3B-Instruct through Novita AI at $0.07/1M input tokens. For the most capable/latest local choice, evaluate Qwen3-Coder-Next with 256k context and reasoning, tool use, function calling, and structured outputs.