Qwen3-30B-A3B-Instruct-2507 is part of the Qwen lineup from Qwen, designed for production workloads that need reliable quality at scale. Instruction-tuned 30B MoE.
It accepts Text inputs and produces Text outputs, with a context window of up to 262,144 input / 16,384 output tokens. The model is tuned for chat, analysis, and tool use in enterprise pipelines.
Balanced pre-training and alignment deliver solid instruction following, summarization, and structured JSON output.
Benchmarks
Benchmark (Area)
Score*
AIME 2024 (Math)
75.4% accuracy (Max)
LiveCodeBench (Coding)
71.4 Pass@1 (Max)
Codeforces Rating (Coding)
2380 rating (Max)
SWE-bench Verified (Coding)
61.6% resolved (Max)
MMLU-Pro (Knowledge)
67.2 EM (Max)
HMMT 2024 Feb (Math)
73.9 Pass@1 (Max)
GPQA Diamond (Science)
68.7 Pass@1 (Max)
MRCR 1M (Long Context)
61.4 MMR (Max)
Terminal Bench 2.0 (Agent)
44.4 Acc (Max)
Apex Shortlist (Agent)
66.8 Pass@1 (Max)
Chinese-SimpleQA (Chinese)
61.5 Pass@1 (Max)
SuperCLUE (Chinese)
53.68 overall (Max)
API request example
Endpoint /v1/chat/completions. Replace <API_KEY> with a key from the console.
Standard OpenAI Chat Completions. Swap `model` for the target slug.