Language modelActive

Gemma 3 27B

Google

Released: March 10, 2025
Data date: September 8, 2026
Predecessor: Gemma 2 27B
Successor: Gemma 4 31B

Price: $0.119 · $0.2
Input · Output / 1M tokens
Input: Text, Image
Output: Text

What was new in Gemma 3 27B compared with Gemma 2 27B was visual input in an open 27B checkpoint. Google reports 14 trillion training tokens, more than 140 training languages, and Gemma 4 31B as the later family.

The checkpoint understands images and generates text. The weight page gives the exact google/gemma-3-27b-it path and documents local use with Transformers and vLLM.

Specifications and access

Specification	Value and source
Weights	Open weightsSource
Access	API, LocalSource
Parameters	27BSource
Context window	128,000 tokensSource
Knowledge cutoff	August 2024Source
Release	2025-03-12Source
Parameters	27 BSource
Input	Text, ImageSource
Output	TextSource
Architecture	Decoder-only transformer with multimodal vision encoder; instruction-tuned variantSource
License	Gemma Terms of UseSource
Training tokens	14,000,000,000,000 tokensSource
Knowledge cutoff	2024-08Source
Languages	140+Source

Measurements without matching peer values

These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.

GPQA Diamond (Show measurement, test conditions, and source)

Source value: 42.4
Score: 42.4
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

SimpleQA (Show measurement, test conditions, and source)

Source value: 10
Score: 10
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

FACTS Grounding (Show measurement, test conditions, and source)

Source value: 74.9
Score: 74.9
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; -; instruction-tuned 27B

Source Google DeepMind

BIG-Bench Hard (Show measurement, test conditions, and source)

Source value: 87.6
Score: 87.6
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

BIG-Bench Extra Hard (Show measurement, test conditions, and source)

Source value: 19.3
Score: 19.3
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

IFEval (Show measurement, test conditions, and source)

Source value: 90.4
Score: 90.4
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

MMLU Pro (Show measurement, test conditions, and source)

Source value: 67.5
Score: 67.5
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

LiveCodeBench (Show measurement, test conditions, and source)

Source value: 29.7
Score: 29.7
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

Bird-SQL dev (Show measurement, test conditions, and source)

Source value: 54.4
Score: 54.4
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; -; instruction-tuned 27B

Source Google DeepMind

Math (Show measurement, test conditions, and source)

Source value: 89
Score: 89
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

HiddenMath (Show measurement, test conditions, and source)

Source value: 60.3
Score: 60.3
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 0-shot; instruction-tuned 27B

Source Google DeepMind

MBPP (Show measurement, test conditions, and source)

Source value: 74.4
Score: 74.4
Metric: percent
Unit: %
Category: vendor-evaluation
Direction: Higher is better
Source type: vendor-reported
Evaluator: Google DeepMind
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. Gemma 3 model-card table; 3-shot; instruction-tuned 27B

Source Google DeepMind

Page 1 of 6

Predecessor and successor

Earlier model generation: Gemma 2 27B
Previous generation in the directory
Model lineage evidence
Google's Gemma 3 model card describes Gemma 3 as "available in more sizes than previous versions"; Gemma 3's 27B variant is the same nominal top dense size as Gemma 2's 27B, but no source names the two 27B checkpoints as directly replacing one another, so the link is recorded as nearest-profile.
Google Gemma 3 model card (2026-09-19)
Later model generation: Gemma 4 31B
Next generation in the directory
Model lineage evidence
Google's Gemma 4 model card frames the new generation explicitly "Compared to Gemma 3," and the Gemma 4 launch post lists 31B Dense as the largest of Gemma 4's four sizes, the same flagship-dense role Gemma 3 27B held; no source pairs the two by exact size, so this is recorded as nearest-profile, not direct.
Google Gemma 4 model card (2026-09-19)

Sources and data date

Type	Evidence and data date
Price	Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs. Source: Pricing · API pricing
Input	Documented modalities. Retrieved September 8, 2026. Source: Model documentation · Context window data (retrieved September 8, 2026) · Gemma 3 model card · Editorial description reviewed October 2, 2026
Output	Documented modalities. Retrieved September 8, 2026. Source: Model documentation · Context window data (retrieved September 8, 2026) · Gemma 3 model card · Editorial description reviewed October 2, 2026
Research status	Research date September 8, 2026. 3 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Not found in the inspected sources	active-parameters-billions Original wording Active parameter count is not separately published for this dense model.
Not found in the inspected sources	api-model-id Original wording Gemma is distributed as open weights; the card does not define a hosted Gemini API model ID.
Additional source	Model metadata · Gemma 3 27B instruction-tuned weights · Editorial description reviewed October 2, 2026
Additional source	Artificial Analysis (retrieved September 8, 2026)
Additional source	Model metadata (retrieved September 8, 2026)
Additional source	OpenRouter: Popularity index (derived), CC BY 4.0
Additional source	Vercel AI Gateway: Popularity index (derived), CC BY 4.0
Additional source	Creative Commons Attribution 4.0
Additional source	Google Gemma 4 model card (2026-09-19)

Gemma 3 27B

Specifications and access

Measurements without matching peer values

Our tests

Pelican on a bicycle

Popularity

Published prices

Predecessor and successor

Sources and data date