← Back to all models
G

Gemma 3 27B

by Google DeepMind
Efficient Free & Open Source Multimodal 🏆 Ranked #41 of 85
75.6
Overall Score
out of 100
About

Google DeepMind's flagship open-source model in the Gemma 3 family, capable of running on a single high-end GPU. Gemma 3 27B supports text and images and delivers strong benchmarks across knowledge, reasoning, and instruction following — making it one of the best self-hostable multimodal models available.

Key Metrics
Context Window
128K
tokens
Avg Response
980
milliseconds
Input Cost
$0.0
per million tokens
Output Cost
$0.0
per million tokens
Arena ELO
1290
Chatbot Arena rating
MT-Bench
8.7
out of 10
Benchmark Scores
MMLU
87.5%
HumanEval
77.2%
MATH
72.0%
GPQA
52.9%
MT-Bench
87.0/10
Capability Profile
Strengths & Limitations
Strengths
✓ Open source ✓ Multimodal (text + images) ✓ Single-GPU capable ✓ 128K context ✓ Strong benchmarks for size
Limitations
⚠ Smaller than frontier models ⚠ Less capable on hardest reasoning ⚠ Limited tool use
Ideal Use Cases
Self-hosted AI Privacy-sensitive deployments Edge computing On-premise chatbots Research
Model Details
Provider Google DeepMind
Released 2025-03-12
Type Free & Open Source
Multimodal Yes
Tier Efficient
Global rank #41 / 85
Pricing (USD)
Input tokens $0.0/M
Output tokens $0.0/M
Per 1,000 tokens ≈ $0.0000 input / $0.0000 output
All Benchmarks
MMLU 87.5%
HumanEval 77.2%
MATH 72.0%
GPQA 52.9%
MT-Bench 8.7/10
Arena ELO 1290
Compare this model View Rankings

You might also consider