Qwen 3 30B-A3B

Overview

Alibaba's Qwen 3 MoE with 30B total and just 3B active parameters, supporting hybrid thinking mode. MMLU 81.4, AIME24 80.4, 100+ languages, Apache 2.0.

When to pick this model

• Fast self-hosted chat that toggles into reasoning when needed
• Multilingual production across 100+ languages
• Workloads needing reasoning quality without the verbosity of dedicated reasoners
• Single 24GB GPU deployments wanting MoE inference speed

VRAM requirements by quantization

Quantization	VRAM required
Q4_K_M (recommended)	19 GB
Q5_K_M	23 GB
Q8_0	35 GB
FP16 (no quantization)	62 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

Published benchmark scores

Benchmark	Score
MMLU (base)	81.38
AIME 2024	80.4

Scores published by the model author or aggregated from public leaderboards. Re-measured monthly by our editorial team.

Strengths

• 3B active parameters keeps inference fast and cheap
• MMLU 81.4 and AIME24 80.4 — strong on both general and reasoning
• Apache 2.0
• Hybrid thinking toggle per request
• 100+ language coverage

Limitations

• ~19GB at Q4 — slightly tight on 16GB cards
• Thinking mode adds latency and token cost
• MoE routing complicates some fine-tuning workflows

Architecture & training

Architecture: MoE 128 experts · 30B/3B active · hybrid thinking

Training: Qwen 3 base.

Quick start

ollama run qwen3:30b-a3b

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.