Skip to main content
Back to Explore

Gemma 4 12B (Reasoning)

Google·Released 2026-06-03
Open SourceMultimodal
Comparison data ready73% coverage12/12 fields directly observedUpdated Jul 18, 2026, 6:01 PM

Pricing

Input

$0.10

per 1M tokens

Output

$0.30

per 1M tokens

Blended

$0.15

per 1M tokens

Cheaper than 78% of models. Median price is $0.70/1M tokens.

Cost Calculator

Tokens per day1M
100K100M

Daily

$0.15

Monthly

$4.50

vs. Similar Models

DeepSeek V3.2 SpecialeQ:+0.2
$0.32+115%
Gemma 4 31B (Non-reasoning)Q:-0.2
$0.20+37%
Grok 4.20 0309 v2 (Non-reasoning)Q:-0.2
$3.00+1900%
Nova 2.0 Pro Preview (medium)Q:-0.2
$3.44+2192%

Performance

126

tokens/sec

Faster than 68% of models

1.35

seconds

Faster than 36% of models

17.20

seconds

Faster than 29% of models

Market Median

94 tok/s

34% faster

Median TTFT

1.07s

26% slower

Throughput/Dollar

841

tok/s per $/1M

Speed Comparison

Gemini 3.1 Pro Preview
126 tok/s-0%
Qwen3 VL 30B A3B (Reasoning)
126 tok/s-0%
Sarvam 105B (high)
126 tok/s-0%

Benchmarks

MMLU-ProNot evaluated
GPQA Diamond
75.3%
HLE
14.8%
LiveCodeBenchNot evaluated
SciCode
38.2%
TerminalBench Hard
18.2%
MATH-500Not evaluated
AIMENot evaluated
AIME 2025Not evaluated
IFBench
73.5%
Long Context Recall
55.3%
Tau2
36.3%
Market AverageTop Score

Open Source

Quick Compare

Similar Models

Compare all 7 models