Benched.ai
View
About
Command Palette
Search for a command to run...
Models
Qwen3 4B (Reasoning)
Alibaba
Reasoning
Open
Context
Release Date
Apr 28, 2025
Window
32,000
Pricing
Per 1M tokens
Input
$0.11
Output
$1.26
Blended 3:1
$0.3975
Capabilities
Speed
105 t/s
Input
Output
Reasoning tokens
Latency
TTFT
1.01 ms
500 token response
24.92 s
Benchmarks
Reasoning
●●○○○
Math
●●●●○
Coding
●○○○○
MMLU Pro
69.6%
GPQA
52.2%
HLE
5.1%
SciCode
3.5%
AIME
65.7%
MATH 500
93.3%
LiveCodeBench
46.5%
HumanEval
90.9%