Vertical Tabs
Get Started
Google DeepMind

Gemma 2 27B

Structured output
Model cardHugging FaceLaunch postWebsite
Gemma 2 27B launch post cover
Context
8K
Released
2y ago
Parameters
27B
Input / 1M tokens
$0.65
Output / 1M tokens
$0.65

Sources:OpenRouter

Benchmarks

DeepSWE is the primary coding benchmark where an exact model evaluation exists. Other sourced evaluations appear below; missing scores are not zero.

Not evaluated in DeepSWE 1.1 as of 2026-09-22. Results are only shown for exact model versions in the official dataset. Official leaderboard.

Other published benchmarks

EvaluationReported score and source
GPQA diamond
36.5%
Epoch AI · gemma-2-27b-it
Epoch AI evaluation
OTIS Mock AIME 2024-2025
1.4%
Epoch AI · gemma-2-27b-it
Epoch AI evaluation

Benchmark compilation: Epoch AI, Capabilities & benchmarking (CC BY attribution). Scores are converted to percentages and model names matched to the Vertical Tabs catalog. Individual evaluations are credited to their original sources. External results may carry separate licences and evaluation configurations; do not interpret scores from different harnesses as directly comparable.

Specifications

Developer
Google DeepMind
Model family
Gemma
Type
Language model
Predecessor
—
Successor
—
Released
June 27, 2024
Model ID
google/gemma-2-27b-it
Openness
Open weights
License
—
Parameters
27B
Input modalities
text
Output modalities
text
Modality
text->text
Native context
8K
Tokenizer
Gemini
Max output tokens
2K
Knowledge cutoff
2024-06-30
Hugging Face
google/gemma-2-27b-it
Source freshness
Endpoints synced Oct 15, 2026, 1:41 PM UTC

Open-weight builds

Ollama variants

Every published Ollama tag matched to this model, including parameter editions, instruction/base variants, and quantized builds.

90 builds
VariantParametersTypeQuantizationSizeContextInput
gemma2:2b-instruct-fp162BinstructFP165.2 GB8Ktext
gemma2:2b-instruct-q2_K2BinstructQ2_K1.2 GB8Ktext
gemma2:2b-instruct-q3_K_L2BinstructQ3_K_L1.6 GB8Ktext
gemma2:2b-instruct-q3_K_M2BinstructQ3_K_M1.5 GB8Ktext
gemma2:2b-instruct-q3_K_S2BinstructQ3_K_S1.4 GB8Ktext
gemma2:2b2BinstructQ4_01.6 GB8Ktext
gemma2:2b-instruct-q4_12BinstructQ4_11.8 GB8Ktext
gemma2:2b-instruct-q4_K_M2BinstructQ4_K_M1.7 GB8Ktext
gemma2:2b-instruct-q4_K_S2BinstructQ4_K_S1.6 GB8Ktext
gemma2:2b-instruct-q5_02BinstructQ5_01.9 GB8Ktext
gemma2:2b-instruct-q5_12BinstructQ5_12 GB8Ktext
gemma2:2b-instruct-q5_K_M2BinstructQ5_K_M1.9 GB8Ktext
gemma2:2b-instruct-q5_K_S2BinstructQ5_K_S1.9 GB8Ktext
gemma2:2b-instruct-q6_K2BinstructQ6_K2.2 GB8Ktext
gemma2:2b-instruct-q8_02BinstructQ8_02.8 GB8Ktext
gemma2:2b-text-fp162BtextFP165.2 GB8Ktext
gemma2:2b-text-q2_K2BtextQ2_K1.2 GB8Ktext
gemma2:2b-text-q3_K_L2BtextQ3_K_L1.6 GB8Ktext
gemma2:2b-text-q3_K_M2BtextQ3_K_M1.5 GB8Ktext
gemma2:2b-text-q3_K_S2BtextQ3_K_S1.4 GB8Ktext
gemma2:2b-text-q4_02BtextQ4_01.6 GB8Ktext
gemma2:2b-text-q4_12BtextQ4_11.8 GB8Ktext
gemma2:2b-text-q4_K_M2BtextQ4_K_M1.7 GB8Ktext
gemma2:2b-text-q4_K_S2BtextQ4_K_S1.6 GB8Ktext
gemma2:2b-text-q5_02BtextQ5_01.9 GB8Ktext
gemma2:2b-text-q5_12BtextQ5_12 GB8Ktext
gemma2:2b-text-q5_K_M2BtextQ5_K_M1.9 GB8Ktext
gemma2:2b-text-q5_K_S2BtextQ5_K_S1.9 GB8Ktext
gemma2:2b-text-q6_K2BtextQ6_K2.2 GB8Ktext
gemma2:2b-text-q8_02BtextQ8_02.8 GB8Ktext
gemma2:9b-instruct-fp169BinstructFP1618 GB8Ktext
gemma2:9b-instruct-q2_K9BinstructQ2_K3.8 GB8Ktext
gemma2:9b-instruct-q3_K_L9BinstructQ3_K_L5.1 GB8Ktext
gemma2:9b-instruct-q3_K_M9BinstructQ3_K_M4.8 GB8Ktext
gemma2:9b-instruct-q3_K_S9BinstructQ3_K_S4.3 GB8Ktext
gemma2:latest9BinstructQ4_05.4 GB8Ktext
gemma2:9b-instruct-q4_19BinstructQ4_16 GB8Ktext
gemma2:9b-instruct-q4_K_M9BinstructQ4_K_M5.8 GB8Ktext
gemma2:9b-instruct-q4_K_S9BinstructQ4_K_S5.5 GB8Ktext
gemma2:9b-instruct-q5_09BinstructQ5_06.5 GB8Ktext
gemma2:9b-instruct-q5_19BinstructQ5_17 GB8Ktext
gemma2:9b-instruct-q5_K_M9BinstructQ5_K_M6.6 GB8Ktext
gemma2:9b-instruct-q5_K_S9BinstructQ5_K_S6.5 GB8Ktext
gemma2:9b-instruct-q6_K9BinstructQ6_K7.6 GB8Ktext
gemma2:9b-instruct-q8_09BinstructQ8_09.8 GB8Ktext
gemma2:9b-text-fp169BtextFP1618 GB8Ktext
gemma2:9b-text-q2_K9BtextQ2_K3.8 GB8Ktext
gemma2:9b-text-q3_K_L9BtextQ3_K_L5.1 GB8Ktext
gemma2:9b-text-q3_K_M9BtextQ3_K_M4.8 GB8Ktext
gemma2:9b-text-q3_K_S9BtextQ3_K_S4.3 GB8Ktext
gemma2:9b-text-q4_09BtextQ4_05.4 GB8Ktext
gemma2:9b-text-q4_19BtextQ4_16 GB8Ktext
gemma2:9b-text-q4_K_M9BtextQ4_K_M5.8 GB8Ktext
gemma2:9b-text-q4_K_S9BtextQ4_K_S5.5 GB8Ktext
gemma2:9b-text-q5_09BtextQ5_06.5 GB8Ktext
gemma2:9b-text-q5_19BtextQ5_17 GB8Ktext
gemma2:9b-text-q5_K_M9BtextQ5_K_M6.6 GB8Ktext
gemma2:9b-text-q5_K_S9BtextQ5_K_S6.5 GB8Ktext
gemma2:9b-text-q6_K9BtextQ6_K7.6 GB8Ktext
gemma2:9b-text-q8_09BtextQ8_09.8 GB8Ktext
gemma2:27b-instruct-fp1627BinstructFP1654 GB8Ktext
gemma2:27b-instruct-q2_K27BinstructQ2_K10 GB8Ktext
gemma2:27b-instruct-q3_K_L27BinstructQ3_K_L15 GB8Ktext
gemma2:27b-instruct-q3_K_M27BinstructQ3_K_M13 GB8Ktext
gemma2:27b-instruct-q3_K_S27BinstructQ3_K_S12 GB8Ktext
gemma2:27b27BinstructQ4_016 GB8Ktext
gemma2:27b-instruct-q4_127BinstructQ4_117 GB8Ktext
gemma2:27b-instruct-q4_K_M27BinstructQ4_K_M17 GB8Ktext
gemma2:27b-instruct-q4_K_S27BinstructQ4_K_S16 GB8Ktext
gemma2:27b-instruct-q5_027BinstructQ5_019 GB8Ktext
gemma2:27b-instruct-q5_127BinstructQ5_121 GB8Ktext
gemma2:27b-instruct-q5_K_M27BinstructQ5_K_M19 GB8Ktext
gemma2:27b-instruct-q5_K_S27BinstructQ5_K_S19 GB8Ktext
gemma2:27b-instruct-q6_K27BinstructQ6_K22 GB8Ktext
gemma2:27b-instruct-q8_027BinstructQ8_029 GB8Ktext
gemma2:27b-text-fp1627BtextFP1654 GB8Ktext
gemma2:27b-text-q2_K27BtextQ2_K10 GB8Ktext
gemma2:27b-text-q3_K_L27BtextQ3_K_L15 GB8Ktext
gemma2:27b-text-q3_K_M27BtextQ3_K_M13 GB8Ktext
gemma2:27b-text-q3_K_S27BtextQ3_K_S12 GB8Ktext
gemma2:27b-text-q4_027BtextQ4_016 GB8Ktext
gemma2:27b-text-q4_127BtextQ4_117 GB8Ktext
gemma2:27b-text-q4_K_M27BtextQ4_K_M17 GB8Ktext
gemma2:27b-text-q4_K_S27BtextQ4_K_S16 GB8Ktext
gemma2:27b-text-q5_027BtextQ5_019 GB8Ktext
gemma2:27b-text-q5_127BtextQ5_121 GB8Ktext
gemma2:27b-text-q5_K_M27BtextQ5_K_M19 GB8Ktext
gemma2:27b-text-q5_K_S27BtextQ5_K_S19 GB8Ktext
gemma2:27b-text-q6_K27BtextQ6_K22 GB8Ktext
gemma2:27b-text-q8_027BtextQ8_029 GB8Ktext

Infrastructure

Provider availability

All providers
ProviderEndpointContextInput /M tokensOutput /M tokensCached input /M tokensCache write /M tokensUptime 1dData handling
NextBitnextbit/int48K$0.65$0.65——100.00%
ZDR: Yes

Training: Unknown

Prompt logging: Unknown

Retention: Unknown

ZDR source Privacy policy

Checked 2026-10-15

Provider data:OpenRouter

Text provenance

Gemini app/web

Google SynthID text watermarking

Google DeepMind says text generated in the Gemini app and web experience is watermarked with SynthID by adjusting token probabilities.

Official explanation

Deployment graph

Can I run this?

Vertical Tabs only shows compatibility when it has structured evidence tying this exact model to hardware, runtime, quantization, memory use, and observed performance.

Browse hardware