Models

Qwen3 30B A3B 2507 Instruct

Qwen3 30B A3B is a 30-billion parameter instruction-tuned model from Alibaba's Qwen3 series, likely utilizing a Mixture-of-Experts architecture with 3 billion active parameters. It is optimized for strong instruction following, reasoning, and multilingual (especially Chinese) performance, balancing capability with inference efficiency.

CodingReasoningFastCheap
Input / 1M tokens
$0.15
Output / 1M tokens
$0.40
Output tokens/s
122.46
First-token seconds
1.12s
Supported plans
0

Benchmark history

Evaluations

15

TAU2

Measured May 14, 2026Source

Score

0.1

Terminalbench Hard

Measured May 14, 2026Source

Score

0.06

Lcr

Measured May 14, 2026Source

Score

0.23

Ifbench

Measured May 14, 2026Source

Score

0.33

Aime 25

Measured May 14, 2026Source

Score

0.66

Aime

Measured May 14, 2026Source

Score

0.73

Math 500

Measured May 14, 2026Source

Score

0.98

Scicode

Measured May 14, 2026Source

Score

0.3

Livecodebench

Measured May 14, 2026Source

Score

0.52

Hle

Measured May 14, 2026Source

Score

0.07

Gpqa

Measured May 14, 2026Source

Score

0.66

Mmlu Pro

Measured May 14, 2026Source

Score

0.78

Artificial Analysis Math Index

Measured May 14, 2026Source

Score

66.3

Artificial Analysis Coding Index

Measured May 14, 2026Source

Score

14.2

Artificial Analysis Intelligence Index

Measured May 14, 2026Source

Score

15

Plan availability

Products and plans that support this model

0
No products or plans have been linked to this model yet.

Discussion

Thinking... Make sure you are connected to GitHub server