---
id: "qwen/qwen3-coder-30b-a3b-instruct"
also: ["qwen/qwen3-coder-30b-a3b-instruct", "Qwen/Qwen3-Coder-30B-A3B-Instruct", "hf/Qwen/Qwen3-Coder-30B-A3B-Instruct"]
cheapest_list_usd_per_m: "0.07 in, 0.27 out, at novita/fp8 on openrouter"
cheapest_effective_usd_per_m: "0.0699 in, 0.2767 out, at siliconflow/fp8 on openrouter"
context: 262144
max_output: 235929
input: ["text"]
output: ["text"]
released: "2025-07-31"
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "logprobs", "max_tokens", "presence_penalty", "repetition_penalty", "response_format", "seed", "stop", "structured_outputs", "temperature", "tool_choice", "tools", "top_k", "top_logprobs", "top_p"]
endpoints: 8
gateways: "openrouter (6), hf (2)"
no_training: 4
trains: 0
retains_prompts: 1
---
# Qwen: Qwen3 Coder 30B A3B Instruct

Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use. Built on the...