Gemma 4 12B
Reasoning
Creator
Google DeepMind
Released
2026-06-03
Intelligence
21.8
Artificial Analysis Index
Coding
31.0
Artificial Analysis Index
In $/1M
$0.10
input tokens
Out $/1M
$0.30
output tokens
Blended $/1M
$0.15
3:1 blended
Speed
24
tokens / sec
Benchmark breakdown
Independent evaluation scores, normalized to 0–100.
GPQA Diamond
75
Humanity’s Last Exam
15
MMLU-Pro—
SciCode
38
LiveCodeBench—
MATH-500—
AIME 2025—
τ²-Bench (agentic)
36
Terminal-Bench Hard
18
IFBench
74