DeepYard
Q

Qwen3 Coder

20

Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation mod...

Open Sourcepaid

About

Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team. It is optimized for agentic coding tasks such as function calling, tool use, and long-context reasoning over repositories. The model features 480 billion total parameters, with 35 billion active per forward pass (8 out of 160 experts). Pricing for the Alibaba endpoints varies by context length. Once a request is greater than 128k input tokens, the higher pricing is used.

Details

Modalitytext
Context Window262.1K tokens
Release DateJul 23, 2025
API AvailableYes
Hostingself-hosted, api
Output Speed100 tokens/sec
Time to First Token700ms
Quality Index84/100
Coding Index90/100
Reasoning Index82/100

Benchmarks

Chatbot Arena ELO
1430/1500
MMLU-Pro
82/100
SWE-bench Verified
70.6/100
MATH-500
88/100
GPQA Diamond
68/100
HumanEval
96/100

Tags

codingopen-weightapichat