Skip to main content
Back to Explore

Qwen3 30B A3B 2507 (Reasoning)

Alibaba·Released 2025-07-30
Open Source

Pricing

Input

$0.28

per 1M tokens

Output

$1.85

per 1M tokens

Blended

$0.67

per 1M tokens

Cheaper than 47% of models. Median price is $0.54/1M tokens.

Cost Calculator

Tokens per day1M
100K100M

Daily

$0.67

Monthly

$20.19

vs. Similar Models

Mistral Large 3Q:+0.1
$0.75+11%
DeepSeek V3 0324Q:-0.1
$1.21+80%
Qwen3.5 4B (Non-reasoning)Q:+0.2
$0.06-91%
INTELLECT-3Q:-0.2
$0.43-37%

Performance

149

tokens/sec

Faster than 74% of models

1.02

seconds

Faster than 55% of models

14.44

seconds

Faster than 34% of models

Market Median

94 tok/s

59% faster

Median TTFT

1.10s

7% faster

Throughput/Dollar

222

tok/s per $/1M

Speed Comparison

Mistral Small 3.2
149 tok/s+0%
Mistral Small 4 (Non-reasoning)
149 tok/s-0%
GPT-5 nano (minimal)
149 tok/s-0%

Benchmarks

MMLU-Pro
80.5%
GPQA Diamond
70.7%
HLE
9.8%
LiveCodeBench
70.7%
SciCode
33.3%
TerminalBench Hard
5.3%
MATH-500
97.6%
AIME
90.7%
AIME 2025
56.3%
IFBench
50.7%
Long Context Recall
59.0%
Tau2
28.1%
Market AverageTop Score

Open Source

Quick Compare

Similar Models

Compare all 7 models