← Models
Model APILLM
Google DeepMind·Apache 2.0
Gemma 4 31B IT
31B · Dense · Instruction-tuned
Largest dense model in the Gemma 4 family — multimodal text and image input, strong coding and reasoning, with a 256K context window.
Model details
| Developed by | Google DeepMind |
| Model family | Gemma 4 |
| Architecture | Dense transformer |
| Parameters | 30.7B |
| Context length | 256K tokens |
| Modalities | Text, Image → Text |
| Vocabulary | 262K |
| License | Apache 2.0 |
On Autoloops
| Hardware | B200 |
| API | OpenAI-compatible |
| Region | US West |
| Pricing | $0.20 / $0.45 / $0.14 per 1M tokens (input / output / cached) |
| Model ID | gemma-4-31b-it |
Capabilities
- Instruction-tuned chat and reasoning with configurable thinking mode
- Native image understanding — OCR, charts, documents, UI screens
- Strong coding, math, and agentic tool-use benchmarks in the Gemma 4 family
- Multilingual support across 140+ pretrain languages
- OpenAI-compatible chat completions via Autoloops
OpenAI-compatible chat completion
from openai import OpenAI
import os
client = OpenAI(
api_key=os.environ["AUTOLOOPS_API_KEY"],
base_url="https://api.autoloops.ai/v1",
)
stream = client.chat.completions.create(
model="gemma-4-31b-it",
messages=[{"role": "user", "content": "Tell me a fun fact about cats."}],
stream=True,
)
for chunk in stream:
print(chunk.choices[0].delta.content or "", end="")