NOUMENOpen app
NOUMEN

Google · Open weights

Gemma 3 12B IT

Added Jan 1, 2024Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Gemma 3 has a large, 128K context window, multilingual support in over 140 languages, and is available in more sizes than previous versions.

  • PDF input
Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.27
Output
$0.27

Reference prices as published by the model provider. API example

Specifications

Context window
131.1K
Max output
16.4K
Average output (7 days)
168 tokens
Longer than 16% of 468 models

Side by side

Gemma 3 12B IT compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Gemma 3 12B ITGoogleThis modelCohere: Command R+cohereRemove
ProviderGooglecohere
Discovery tierOpen weightsBalanced
AddedJan 1, 2024Jan 1, 2024
Limits
Context window131.07KBest128K
Max output16.38KBest4.1K
Average output168 tokens146 tokens
Pricing, USD per 1M tokens
Input$0.27Best$2.86
Output$0.27Best$14.25
Performance
Request success (24h)100.0%—
Benchmark indexes
Intelligence Index3.85.3Best
Coding Index5.8—
Math Index18.3—
Capabilities
PDF inputYesNo
Reasoning effort——

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Gemma 3 12B IT
  • Cohere: Command R+

Intelligence Index

  • Gemma 3 12B IT3.8
  • Cohere: Command R+5.3

Gemma 3 12B IT: Better than 0% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Gemma 3 12B IT: 5.8
    • Cohere: Command R+: —

  • Math Index

    Composite of the math evaluations

    • Gemma 3 12B IT: 18.3
    • Cohere: Command R+: —

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Gemma 3 12B IT: 0.2%
    • Cohere: Command R+: —

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Gemma 3 12B IT: 0.0%
    • Cohere: Command R+: —

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Gemma 3 12B IT: 1.0%
    • Cohere: Command R+: —

  • AA-LCRv1.1

    Reasoning across long documents

    • Gemma 3 12B IT: 8.3%
    • Cohere: Command R+: —

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Gemma 3 12B IT: 4.2%
    • Cohere: Command R+: 4.6%Top

  • GPQA Diamond

    Graduate-level science questions

    • Gemma 3 12B IT: 34.9%Top
    • Cohere: Command R+: 32.3%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • Gemma 3 12B IT: 59.5%Top
    • Cohere: Command R+: 43.2%

  • IFBench

    Following precise output instructions

    • Gemma 3 12B IT: 36.7%
    • Cohere: Command R+: —

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Gemma 3 12B IT: 0.0%
    • Cohere: Command R+: —

  • SciCode

    Scientific computing in Python

    • Gemma 3 12B IT: 16.4%
    • Cohere: Command R+: —

  • LiveCodeBench

    Recent competitive programming problems

    • Gemma 3 12B IT: 13.7%Top
    • Cohere: Command R+: 12.2%

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • Gemma 3 12B IT: 18.3%
    • Cohere: Command R+: —

  • AIME

    Competition mathematics

    • Gemma 3 12B IT: 22.0%Top
    • Cohere: Command R+: 0.7%

  • MATH-500

    Competition math problems

    • Gemma 3 12B IT: 85.3%Top
    • Cohere: Command R+: 27.9%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Gemma 3 12B IT: 0.8%
    • Cohere: Command R+: —

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Gemma 3 12B IT: 10.8%
    • Cohere: Command R+: —

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Gemma 3 12B IT: 8.3%
    • Cohere: Command R+: —

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Request success (24h)
100.0%
Reliable
Hourly request success · last 72 hoursLowest hour 100.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "unsloth/gemma-3-12b-it",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="unsloth/gemma-3-12b-it",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Google

Other Google models, starting with the same family.