modelbenchmark.io

Gemma 3 12B

Google · gemma-3-12b

Compare

Google's open-weight Gemma 3 vision-language model for text and image understanding

Specification

most-agreed values

Context
131K
Max output
8K
Released
2025-03-13
Knowledge cutoff
2024-08-31
Retires
Open weights
yes
Input
text, image
Output
text

Price

US dollars per million tokens · most-agreed

Input
$0.15
Output
$0.50
Cache read
$0.15
Cache write
$0.15

Quality

3 benchmarks · 1 source

Composite
18th
percentile of 334 scored models
Rank
274 / 334
±0.10 sd
Evidence
3 × 1
one source, or fewer than 3 benchmarks
Effort range
one configuration only
reasoning18th2/4 bench
math15th1/7 bench

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
Chess Puzzles0.0 ±0.0defaultEpoch AI2026-08-28
OTIS Mock AIME 2024-202516.7 ±4.2defaultEpoch AI2026-08-28
GPQA diamond39.5 ±2.5defaultEpoch AI2026-08-28

The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.

Available from 2 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
databricks$0.15$0.50$0.15$0.15128K32K
neon$0.15$0.50131K8K

More from Google

most-hosted first