Skip to content
ShilpBot
AI Models

Google Gemma 4 31B (Free): Specs, Limits and How to Use It

Google Gemma 4 31B free: dense multimodal instruction-tuned model, image/text/video input, 262K context. Specs, limits, test prompts and how to try it.

By ShilpBot Team · Published · 3 min read

Gemma 4 31B Instruct is Google DeepMind's dense multimodal model with about 30.7 billion parameters. Its listing mentions a 256K-token context window, a configurable thinking mode and native function calling.

Unlike the mixture-of-experts 26B sibling, this model uses all of its parameters for every token. That usually means more consistent quality per request at the cost of more compute.

Specifications checked against the public OpenRouter catalogue on 2026-11-16. See the live model page for current limits.

Google Gemma 4 31B quick facts

DetailValue
DeveloperGoogle DeepMind
Model IDgoogle/gemma-4-31b-it:free
Added to the catalogue2026-04-02
TypeDense multimodal instruction-tuned model
Parameters30.7B
Context window262K tokens (262,144)
Maximum output32,768 tokens
Input typesimage, text, video
PriceFree (paid listing: $0.20 in / $0.40 out per 1M tokens)

Feature support from the catalogue: Tool calling: listed; JSON / structured output: listed; Reasoning: supported, effort not adjustable.

What Google Gemma 4 31B is good for

  • General-purpose assistant work where you want consistent answers.
  • Function calling and tool use, which the listing highlights.
  • Multimodal tasks that mix text, images and video.

Limits of Google Gemma 4 31B

  • Dense models are slower per token than similar-quality mixture-of-experts models.
  • Output is capped at 32,768 tokens.
  • Free tiers can be rate limited during busy periods.

How Google Gemma 4 31B compares

  • Google Gemma 4 26B A4B: Both have a 262,144-token context window; both cap output at 32,768 tokens.
  • NVIDIA Nemotron 3.5 Super VL: It has a smaller context window (262,144 against 1,000,000 tokens); its output cap is 32,768 tokens against 131,072.
  • inclusionAI Ling 3.2 Flash: It has a smaller context window (262,144 against 1,048,576 tokens); its output cap is 32,768 tokens against 262,144; it also accepts video input.

Calling Google Gemma 4 31B from code

Use the model ID google/gemma-4-31b-it:free with any OpenAI-compatible client pointed at OpenRouter:

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "google/gemma-4-31b-it:free", "messages": [{"role": "user", "content": "Explain what you can do in two sentences."}]}'

Prompts that show what Google Gemma 4 31B can do

  1. Ask it to call a made-up weather tool with the right arguments for five different questions.
  2. Give it a chart image and ask for the key takeaway and one caveat.
  3. Turn thinking mode on and off for a reasoning question and compare the answers.

What Google Gemma 4 31B costs

The free listing costs nothing, but the same model also has a paid listing at about $0.20 per million input tokens and $0.40 per million output tokens. As a rough guide, sending 1 million input tokens and receiving 100,000 output tokens would cost about $0.24 on the paid listing. Use the free tier to test, and move to the paid route if you need steadier availability. Free pricing is set by the provider and can change at any time, so check the model page before you plan around it.

Keep reading

All free models are compared in our free AI chat models roundup. New to terms like mixture-of-experts? Start with how neural networks learn.

Frequently asked questions

Should I choose Gemma 4 31B or 26B A4B?

Try 26B A4B first for speed and cost. Move to 31B if you need more consistent quality, and test both on your own prompts.

Is Google Gemma 4 31B free, and what are its limits?

As of 2026-11-16 the catalogue lists it at zero price. 262K tokens (262,144) context window, up to 32,768 output tokens, model ID google/gemma-4-31b-it:free. Free use can be rate limited and providers may change it.

#Google Gemma 4 31B#Google DeepMind free model#free AI model#AI chat model#general model#Kilo Gateway#OpenRouter