Gemma 3 27B is the flagship multimodal model in Google DeepMind's Gemma 3 family, designed to tackle the most complex challenges in artificial intelligence. With 27 billion parameters and a context window of up to 128K tokens, the model combines high performance with computational efficiency, delivering state-of-the-art results in both text-based and multimodal tasks.
The model is built upon an architecture inherited from Gemini 2.0 and employs advanced distillation techniques along with an optimized attention structure. This allows it to run even on a single GPU while maintaining performance comparable to models that require up to 32 GPUs. Thanks to its extended context window, SigLIP encoder, and "Pan&Scan" technology, the model can process long documents, scientific papers, and images within a single query without losing coherence or response quality. This makes it particularly valuable for legal and technical analysis, medical diagnostics, and content research involving visual elements. Gemma 3 27B supports more than 140 languages, and an improved tokenizer ensures high-quality translation and cross-lingual analysis.
Gemma 3 27B is not just another language model—it is a next-generation AI assistant that combines the power of large-scale models with accessibility and flexibility. Among its peers, it stands out due to its unique combination of efficiency, multimodality, multilingual support, and high performance on a single graphics card.
| Model Name | Context | Type | GPU | TPS | Status | Link |
|---|---|---|---|---|---|---|
| RedHatAI/gemma-3-27b-it-quantized.w4a16 | 128,000.0 | Public | 2×RTX4090 tensor |
AVAILABLE | chat |
curl https://chat.immers.cloud/v1/endpoints/gemma3-27b-int4/generate/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer USER_API_KEY" \
-d '{"model": "Gemma-3-27b-it", "messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Say this is a test"}
], "temperature": 0, "max_tokens": 150}'
$response = Invoke-WebRequest https://chat.immers.cloud/v1/endpoints/gemma3-27b-int4/generate/chat/completions `
-Method POST `
-Headers @{
"Authorization" = "Bearer USER_API_KEY"
"Content-Type" = "application/json"
} `
-Body (@{
model = "Gemma-3-27b-it"
messages = @(
@{ role = "system"; content = "You are a helpful assistant." },
@{ role = "user"; content = "Say this is a test" }
)
} | ConvertTo-Json)
($response.Content | ConvertFrom-Json).choices[0].message.content
#!pip install OpenAI --upgrade
from openai import OpenAI
client = OpenAI(
api_key="USER_API_KEY",
base_url="https://chat.immers.cloud/v1/endpoints/gemma3-27b-int4/generate/",
)
chat_response = client.chat.completions.create(
model="Gemma-3-27b-it",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Say this is a test"},
]
)
print(chat_response.choices[0].message.content)
Rent your own physically dedicated instance with hourly or long-term monthly billing.
We recommend deploying private instances in the following scenarios:
| Name | GPU | TPS | Max Concurrency | |||
|---|---|---|---|---|---|---|
131,072.0 pipeline |
3 | $0.88 | 1.172 | Launch | ||
131,072.0 tensor |
2 | $0.93 | 39.820 | 1.355 | Launch | |
131,072.0 tensor |
4 | $0.96 | 2.043 | Launch | ||
131,072.0 pipeline |
3 | $1.06 | 1.172 | Launch | ||
131,072.0 tensor |
2 | $1.23 | 44.970 | 1.355 | Launch | |
131,072.0 tensor |
4 | $1.26 | 2.043 | Launch | ||
131,072.0 tensor |
2 | $1.56 | 38.370 | 1.355 | Launch | |
131,072.0 tensor |
2 | $1.92 | 52.790 | 1.355 | Launch | |
131,072.0 tensor |
2 | $2.22 | 2.409 | Launch | ||
131,072.0 |
1 | $2.37 | 51.560 | 3.647 | Launch | |
131,072.0 tensor |
2 | $2.93 | 53.100 | 2.409 | Launch | |
131,072.0 |
1 | $3.83 | 49.870 | 3.647 | Launch | |
131,072.0 |
1 | $4.11 | 63.370 | 4.569 | Launch | |
131,072.0 |
1 | $4.74 | 7.667 | Launch | ||
| Name | GPU | TPS | Max Concurrency | |||
|---|---|---|---|---|---|---|
131,072.0 tensor |
4 | $0.96 | 1.644 | Launch | ||
131,072.0 tensor |
4 | $1.26 | 1.644 | Launch | ||
131,072.0 pipeline |
3 | $1.34 | 2.355 | Launch | ||
131,072.0 tensor |
4 | $1.57 | 3.753 | Launch | ||
131,072.0 tensor |
2 | $2.22 | 2.010 | Launch | ||
131,072.0 pipeline |
3 | $2.29 | 2.355 | Launch | ||
131,072.0 tensor |
4 | $2.34 | 3.753 | Launch | ||
131,072.0 |
1 | $2.37 | 3.248 | Launch | ||
131,072.0 pipeline |
3 | $2.83 | 2.355 | Launch | ||
131,072.0 tensor |
4 | $2.89 | 3.753 | Launch | ||
131,072.0 tensor |
2 | $2.93 | 2.010 | Launch | ||
131,072.0 tensor |
4 | $3.60 | 3.753 | Launch | ||
131,072.0 |
1 | $3.83 | 3.248 | Launch | ||
131,072.0 |
1 | $4.11 | 4.171 | Launch | ||
131,072.0 |
1 | $4.74 | 7.268 | Launch | ||
| Name | GPU | TPS | Max Concurrency | |||
|---|---|---|---|---|---|---|
131,072.0 pipeline |
6 | $1.65 | 1.249 | Launch | ||
131,072.0 tensor |
4 | $1.75 | 1.615 | Launch | ||
131,072.0 tensor |
4 | $2.34 | 1.615 | Launch | ||
131,072.0 |
1 | $2.50 | 23.640 | 1.109 | Launch | |
131,072.0 tensor |
4 | $2.97 | 29.890 | 1.615 | Launch | |
131,072.0 tensor |
4 | $3.68 | 40.920 | 1.615 | Launch | |
131,072.0 pipeline |
3 | $3.89 | 1.798 | Launch | ||
131,072.0 |
1 | $3.95 | 26.520 | 1.109 | Launch | |
131,072.0 |
1 | $4.11 | 2.032 | Launch | ||
131,072.0 pipeline |
3 | $4.34 | 1.798 | Launch | ||
131,072.0 tensor |
4 | $4.35 | 3.724 | Launch | ||
131,072.0 |
1 | $4.74 | 5.130 | Launch | ||
131,072.0 tensor |
4 | $5.74 | 3.724 | Launch | ||
Contact our dedicated neural networks support team at nn@immers.cloud or send your request to the sales department at sale@immers.cloud.