GeekRadar / The comparison desk

COMPARE.

AI models

google

Gemma 3 12B IT

AI model Released Mar 12, 2025

API mode: Chat^*
Context window: 131,072 tokens
Maximum output: 131,072 tokens
Open weights: Yes

Capability & preference

Human preference · Arena

Overall text preferences from head-to-head votes. Not factual accuracy or a task-success rate.

Gemma 3 12B IT 1334.2 Gemma 3 12B IT · 95% interval 1324.8–1343.6 · 3,829 votes · 2026-10-28

Model performance

Radar task score: four tests, equal weight. An experimental summary of these tasks, not a universal intelligence rating.

Gemma 3 12B IT

Tested configuration Radar task score

Unrated

1/4 core tests · not enough evidence

2 benchmark results for this configuration

Scientific reasoning

GPQA Diamond · 25%

Expert-written science questions; strict multiple-choice scoring. Chance accuracy is 25%.

Gemma 3 12B IT 39.5^*% Approx. 95% interval: 34.5–44.4% · 2026-08-28

Novel problem solving

Mystery Game Puzzles · 25%

Best-next-move puzzles in an undisclosed game variant. A narrow reasoning test, not general intelligence.

Gemma 3 12B IT Not available

Advanced mathematics

FrontierMath Tiers 1-3 · 25%

Private v2 mathematics problems with Python tools. Earlier versions and Tier 4 are excluded.

Gemma 3 12B IT Not available

Factual answers

SimpleQA Verified · 25%

Correct factual answers under a guessing prompt, graded by a model. Not a hallucination or abstention score.

Gemma 3 12B IT Not available

Competition mathematics

OTIS Mock AIME 2024-2025 · Separate result

Competition problems with model-extracted answers checked against an exact key. Earlier grading procedures are excluded.

Gemma 3 12B IT 16.7^*% Approx. 95% interval: 8.4–25.0% · 2026-08-28

Bars show reported task scores. Whiskers are approximate 95% intervals where standard errors are available. Different reasoning budgets are not equal-cost tests. Missing results are never zero.

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

Test version

Arena text preference / codinghuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,280.88 rating

Confidence interval: 1,258.12 – 1,303.64

543 votes

Oct 28, 2026

Category: coding
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,331.02 rating

Confidence interval: 1,307.99 – 1,354.06

600 votes

Oct 28, 2026

Category: creative_writing
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / englishhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,348.49 rating

Confidence interval: 1,336.72 – 1,360.26

2330 votes

Oct 28, 2026

Category: english
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,289.00 rating

Confidence interval: 1,274.93 – 1,303.07

2526 votes

Oct 28, 2026

Category: exclude_ties
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / experthuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,247.55 rating

Confidence interval: 1,207.18 – 1,287.92

186 votes

Oct 28, 2026

Category: expert
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / germanhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,370.37 rating

Confidence interval: 1,327.98 – 1,412.75

166 votes

Oct 28, 2026

Category: german
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,309.25 rating

Confidence interval: 1,291.70 – 1,326.80

977 votes

Oct 28, 2026

Category: hard_prompts
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,318.18 rating

Confidence interval: 1,297.62 – 1,338.74

700 votes

Oct 28, 2026

Category: hard_prompts_english
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,376.35 rating

Confidence interval: 1,346.48 – 1,406.22

364 votes

Oct 28, 2026

Category: industry_business_and_management_and_financial_operations
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,297.45 rating

Confidence interval: 1,276.46 – 1,318.44

697 votes

Oct 28, 2026

Category: industry_entertainment_and_sports_and_media
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,380.85 rating

Confidence interval: 1,342.67 – 1,419.02

217 votes

Oct 28, 2026

Category: industry_legal_and_government
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,365.82 rating

Confidence interval: 1,345.32 – 1,386.32

760 votes

Oct 28, 2026

Category: industry_life_and_physical_and_social_science
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,335.88 rating

Confidence interval: 1,305.92 – 1,365.84

333 votes

Oct 28, 2026

Category: industry_mathematical
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,334.15 rating

Confidence interval: 1,287.47 – 1,380.84

155 votes

Oct 28, 2026

Category: industry_medicine_and_healthcare
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,312.80 rating

Confidence interval: 1,294.11 – 1,331.48

867 votes

Oct 28, 2026

Category: industry_software_and_it_services
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,307.02 rating

Confidence interval: 1,289.56 – 1,324.47

1010 votes

Oct 28, 2026

Category: industry_writing_and_literature_and_language
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,299.09 rating

Confidence interval: 1,283.02 – 1,315.16

1145 votes

Oct 28, 2026

Category: instruction_following
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,316.67 rating

Confidence interval: 1,288.74 – 1,344.60

371 votes

Oct 28, 2026

Category: longer_query
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / mathhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,307.40 rating

Confidence interval: 1,280.55 – 1,334.25

389 votes

Oct 28, 2026

Category: math
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,333.71 rating

Confidence interval: 1,308.08 – 1,359.34

471 votes

Oct 28, 2026

Category: multi_turn
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / non_englishhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,317.83 rating

Confidence interval: 1,303.67 – 1,331.98

1499 votes

Oct 28, 2026

Category: non_english
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / overallhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,334.19 rating

Confidence interval: 1,324.82 – 1,343.57

3829 votes

Oct 28, 2026

Category: overall
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / russianhuman preference · 2026-10-28 · rating

Gemma 3 12B IT

1,335.21 rating

Confidence interval: 1,304.06 – 1,366.36

285 votes

Oct 28, 2026

Category: russian
Evaluated configuration: Gemma 3 12B IT
Evaluation type: human preference

Arena text preference / codinghuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,280.98 rating

Confidence interval: 1,258.21 – 1,303.74

543 votes

Oct 15, 2026

Category: coding
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / creative_writinghuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,331.20 rating

Confidence interval: 1,308.16 – 1,354.24

600 votes

Oct 15, 2026

Category: creative_writing
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / englishhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,348.46 rating

Confidence interval: 1,336.69 – 1,360.23

2330 votes

Oct 15, 2026

Category: english
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,289.04 rating

Confidence interval: 1,274.97 – 1,303.11

2526 votes

Oct 15, 2026

Category: exclude_ties
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / experthuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,247.57 rating

Confidence interval: 1,207.18 – 1,287.95

186 votes

Oct 15, 2026

Category: expert
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / germanhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,370.34 rating

Confidence interval: 1,327.96 – 1,412.71

166 votes

Oct 15, 2026

Category: german
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / hard_promptshuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,309.26 rating

Confidence interval: 1,291.71 – 1,326.81

977 votes

Oct 15, 2026

Category: hard_prompts
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,318.08 rating

Confidence interval: 1,297.52 – 1,338.65

700 votes

Oct 15, 2026

Category: hard_prompts_english
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,376.42 rating

Confidence interval: 1,346.54 – 1,406.30

364 votes

Oct 15, 2026

Category: industry_business_and_management_and_financial_operations
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,297.50 rating

Confidence interval: 1,276.51 – 1,318.50

697 votes

Oct 15, 2026

Category: industry_entertainment_and_sports_and_media
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,380.59 rating

Confidence interval: 1,342.40 – 1,418.78

217 votes

Oct 15, 2026

Category: industry_legal_and_government
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,365.85 rating

Confidence interval: 1,345.34 – 1,386.35

760 votes

Oct 15, 2026

Category: industry_life_and_physical_and_social_science
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_mathematicalhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,335.91 rating

Confidence interval: 1,305.94 – 1,365.87

333 votes

Oct 15, 2026

Category: industry_mathematical
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,334.21 rating

Confidence interval: 1,287.50 – 1,380.91

155 votes

Oct 15, 2026

Category: industry_medicine_and_healthcare
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,312.88 rating

Confidence interval: 1,294.19 – 1,331.57

867 votes

Oct 15, 2026

Category: industry_software_and_it_services
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,307.08 rating

Confidence interval: 1,289.62 – 1,324.53

1010 votes

Oct 15, 2026

Category: industry_writing_and_literature_and_language
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / instruction_followinghuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,299.08 rating

Confidence interval: 1,283.01 – 1,315.15

1145 votes

Oct 15, 2026

Category: instruction_following
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / longer_queryhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,316.73 rating

Confidence interval: 1,288.80 – 1,344.67

371 votes

Oct 15, 2026

Category: longer_query
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / mathhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,307.42 rating

Confidence interval: 1,280.57 – 1,334.28

389 votes

Oct 15, 2026

Category: math
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / multi_turnhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,333.75 rating

Confidence interval: 1,308.11 – 1,359.39

471 votes

Oct 15, 2026

Category: multi_turn
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / non_englishhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,317.93 rating

Confidence interval: 1,303.77 – 1,332.08

1499 votes

Oct 15, 2026

Category: non_english
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / overallhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,334.20 rating

Confidence interval: 1,324.83 – 1,343.57

3829 votes

Oct 15, 2026

Category: overall
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Arena text preference / russianhuman preference · 2026-10-15 · rating

Gemma 3 12B IT

1,335.33 rating

Confidence interval: 1,304.17 – 1,366.49

285 votes

Oct 15, 2026

Category: russian
Evaluated configuration: gemma-3-12b-it
Evaluation type: human preference

Run it locally

Inference · memory estimate

Gemma 3 12B IT

22.7 GiB checkpoint tensor payload · 12.19 B stored tensor elements · BF16

Stored tensors can include buffers and repeated shared weights. A verified unique parameter count is unavailable. Checkpoint size is not the memory required to run the model.

API cost

USD / month

Input tokens / monthOutput tokens / month

Standard text-token estimate. Excludes taxes, caching, batch discounts, reasoning and multimodal surcharges. Provider limits and regional availability may differ.

Gemma 3 12B IT

Provider offer

Gemma 3 12B IT / Pricing tiers 1

Provider / scope	Rate	Currency	Amount / unit
DeepInfragoogle/gemma-3-12b-itStandard	Input	USD	0.05 / 1M tokens
DeepInfragoogle/gemma-3-12b-itStandard	Output	USD	0.15 / 1M tokens

Capabilities

Context window: 131,072 tokens
Maximum input: 128,000 tokens^*
Maximum output: 131,072 tokens
Reasoning: No
Tool calling: Yes
Structured output: Yes^*
Open weights: Yes
Knowledge cutoff: 2024-08
API mode: Chat^*
Image input: Yes^*
Audio input: No^*
System messages: Yes^*
Model license: Gemma
Checkpoint tensor payload: 22.70 GiB

Input modalities

text
image

Output modalities

text

System configuration

Detailed specifications

Specifications · 19 fields

deprecation date

2027-03-30^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

input cost per token

9.0e-8^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

input cost per token batches

5.0e-8^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

input cost per token flex

5.0e-8^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

input cost per token priority

1.6e-7^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

litellm provider

bedrock_converse^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

max input tokens

128,000^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

max output tokens

8,000^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

max tokens

8,000^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

mode

chat^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

output cost per token

2.9e-7^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

output cost per token batches

1.5e-7^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

output cost per token flex

1.5e-7^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

output cost per token priority

5.1e-7^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

supports audio input

No^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

supports function calling

Yes^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

supports response schema

Yes^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

supports system messages

Yes^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

supports vision

Yes^*

Scope and conditions

native_context

provider: bedrock_converse^*
mode: chat^*
full_native_model_id: google.gemma-3-12b-it^*
pricing_scope: Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.^*

* Provisional data