Apertus 8B Instruct
Creator
Swiss AI Initiative
Released
2025-09-02
Intelligence
5.9
Artificial Analysis Index
Coding
1.4
Artificial Analysis Index
In $/1M
$0.10
input tokens
Out $/1M
$0.20
output tokens
Blended $/1M
$0.13
3:1 blended
Speed
0
tokens / sec
Benchmark breakdown
Independent evaluation scores, normalized to 0–100.
GPQA Diamond
26
Humanity’s Last Exam
5
MMLU-Pro—
SciCode
4
LiveCodeBench—
MATH-500—
AIME 2025—
τ²-Bench (agentic)
11
Terminal-Bench Hard
0
IFBench
22