DataLearner logo
QW

Qwen3-30B-A3B-2507

Chat modelQwen3

Qwen3-30B-A3B-Instruct-2507

Release date: 2025-07-29Updated: 2025-07-30Views: 1,519
Parameters
30.5B
Context length
256K
Multilingual
Supported
Reasoning ability
2/5

Qwen3-30B-A3B-Instruct-2507 is an AI model published by Alibaba, released on 2025-07-29, for Chat model, with 30.5B parameters, and 256K context length, with a 78.40 score on MMLU Pro.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Qwen3-30B-A3B-2507

Model basics

Reasoning traces
No data
Thinking modes
Standard Mode (Default)Thinking Mode
Context length
256K tokens
Max output length
16K tokens
Model type
Chat model
Modality (in / out)
Text → Text
Release date
2025-07-29
Model file size
No data
MoE architecture
Yes
Total params / Active params
30.5B / 3.3B
Knowledge cutoff
No data
Qwen3-30B-A3B-2507

Open source & experience

Code license
Weights license
Apache 2.0- Commercial use permitted
Live demo
Qwen3-30B-A3B-2507

Official resources

Paper
N/A
DataLearnerAI blog
N/A
Qwen3-30B-A3B-2507

API details

API speed
4/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Standard
TypeConditionInputOutput
Text-$0.500/ 1M tokens$3.00/ 1M tokens
Cache PricingPrompt Cache
TypeTTLWriteRead
Text-$0.625/ 1M tokens$0.050/ 1M tokens

“—” means the modality is not billed in that direction, or the vendor has not published a price for it.

Qwen3-30B-A3B-2507

Benchmark Results

Qwen3-30B-A3B-2507 currently shows benchmark results led by LiveCodeBench (86 / 250, score 70.70), MMLU Pro (72 / 176, score 78.40), IF Bench (154 / 282, score 50.70). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

4 evaluations
Benchmark / mode
Score
Rank/total
MMLU Pro
Standard Mode
78.40
72 / 176
HLE
Thinking Mode
10.30
388 / 563
HLE
Thinking Mode
9.80
392 / 563
CritPt
Thinking Mode
0.30
182 / 200

General Evaluation

2 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Standard Mode
70.40
320 / 462
GPQA Diamond
Thinking Mode
70.70
318 / 462

Coding and Software Engineer

4 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
Standard Mode
43.20
181 / 250
LiveCodeBench
Thinking Mode
70.70
86 / 250
SciCode
Thinking Mode
33
120 / 130
SWE-bench Verified
Thinking Mode
22
115 / 116

Math and Reasoning

2 evaluations
Benchmark / mode
Score
Rank/total
AIME2025
Standard Mode
61.30
137 / 215
AIME2025
Thinking Mode
56.30
143 / 215

Agent Level Benchmark

4 evaluations
Benchmark / mode
Score
Rank/total
τ²-Bench
Thinking ModeTools
49
36 / 44
τ²-Bench - Telecom
Thinking ModeTools
28.10
220 / 264
τ³-Banking
Thinking ModeTools
5.40
149 / 164
Terminal Bench Hard
Thinking ModeTools
5.30
210 / 244

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
Thinking Mode
50.70
154 / 282

AI Agent - Information Search

1 evaluations
Benchmark / mode
Score
Rank/total
BrowseComp
Thinking ModeTools
2.29
57 / 57

AI Agent - Tool Usage

1 evaluations
Benchmark / mode
Score
Rank/total
Terminal-Bench 2.1
Thinking ModeTools
1.50
190 / 192

Multimodal Understanding

1 evaluations
Benchmark / mode
Score
Rank/total
GDP.pdf
Thinking Mode
0.40
118 / 118
Qwen3-30B-A3B-2507

Model variants & downloads

Variant nameVersion typeQuantizationModel sizeHuggingFace link
Qwen3-30B-A3B-Thinking-2507ℹ️InstructFP1661.1 GBDownload link
Qwen3-30B-A3B-Instruct-2507-FP8ℹ️InstructFP831.17GBDownload link
Qwen3-30B-A3B-2507

Publisher

Qwen3-30B-A3B-Instruct-2507

Model Overview

Qwen3-30B-A3B-Instruct-2507 is an AI model published by Alibaba, released on 2025-07-29, for Chat model, with 30.5B parameters, and 256K context length, with a 78.40 score on MMLU Pro.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code