Llama 3.2 Instruct 90B
Vision
Creator
Meta AI (FAIR)
Released
2024-09-25
Intelligence
11.9
Artificial Analysis Index
Coding
—
Artificial Analysis Index
In $/1M
$1.38
input tokens
Out $/1M
$1.38
output tokens
Blended $/1M
$1.38
3:1 blended
Speed
48
tokens / sec
Benchmark breakdown
Independent evaluation scores, normalized to 0–100.
GPQA Diamond
43
Humanity’s Last Exam
5
MMLU-Pro
67
SciCode
24
LiveCodeBench
21
MATH-500
63
AIME 2025—
τ²-Bench (agentic)—
Terminal-Bench Hard—
IFBench—