Gemma 3 12B
by Google · google/gemma-3-12b-it · #22 cheapest of 325 paid models
Prices updated Jul 22, 2026, 1:38 AM UTC · refreshed hourly
Gemma 3 introduces multimodality, supporting vision-language input and text outputs. It handles context windows up to 128k tokens, understands over 140 languages, and offers improved math, reasoning, and chat capabilities,...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$0.05
per 1M tokens
Output
$0.15
per 1M tokens
Blended (3:1)
$0.075
3 input : 1 output
Cache read
—
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for Gemma 3 12B (no cache-read rate published — hits billed as normal input).
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $0.075 | $0.00525 | — |
| 50% | $0.075 | $0.00525 | −0% |
| 90% | $0.075 | $0.00525 | −0% |
*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates for this provider’s batch API. Confirm on the vendor’s pricing page.
What a request costs
| Workload | Input tokens | Output tokens | Cost / request | Cost / 1K requests |
|---|---|---|---|---|
| Short chat message | 500 | 300 | $7.00e-5 | $0.07 |
| Document summary | 8,000 | 1,000 | $0.00055 | $0.55 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.0018 | $1.80 |
| Long-context analysis | 150,000 | 5,000 | $0.00825 | $8.25 |
Performance
Intelligence Index
5.5
AA composite quality
Coding Index
5.8
Math Index
18.3
Agentic Index
0.3
Output speed
0 tok/s
median
Time to first token
0s
median
Time to first answer
0s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| MMLU-Pro | 59.5% |
| GPQA Diamond | 34.9% |
| Humanity's Last Exam | 4.8% |
| LiveCodeBench | 13.7% |
| SciCode | 17.4% |
| MATH-500 | 85.3% |
| AIME | 22.0% |
| AIME 2025 | 18.3% |
| IFBench | 36.7% |
| AA-LCR (Long Context Reasoning) | 6.7% |
| Terminal-Bench Hard | 0.8% |
| Terminal-Bench 2.1 | 0.0% |
| τ²-Bench (Telecom) | 10.8% |
| τ³-Bench Banking | 0.8% |
Specs
Context window
131K
tokens
Max output
16K
tokens
Input modalities
text, image
Tokenizer
Gemini