G
Gemma 4 31BOPEN
Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output.
Prices update automatically — checked daily against provider list prices.
- $0.25 per 1M tokens, assuming 3 input for 1 output the convention used by Artificial Analysis
- cheaper than 56% of the 200 open-weight models tracked here
- 8 of 35 Google models tracked here, from cheapest
Benchmarks & performance
Independent benchmark scores for Gemma 4 31B, measured by Artificial Analysis. Higher is better. Measurement mode: Reasoning.
About this model
Gemma 4 31B is an open-weight AI model by Google. You can download and self-host it for free; the prices below are hosted-API list prices, tracked daily, for when you prefer convenience over self-hosting.
Frequently asked questions
What is Gemma 4 31B?
Gemma 4 31B is an AI language model from Google. It is open-weight: you can download it and run it on your own hardware, for free.
Is Gemma 4 31B free?
The weights are free and open — you can self-host Gemma 4 31B and pay nothing per token. If you prefer a hosted API, list prices are $0.2 per million input tokens and $0.4 per million output tokens.
What is Gemma 4 31B good at?
Independent benchmarks from Artificial Analysis give it GPQA 85.7%, Humanity's Last Exam 23.6%, Long Context Reasoning 69.7%, SciCode 45.5%, IFBench 75.6%, τ²-Bench 59.9%, τ-Bench Banking 14.8%, Terminal-Bench 43.4%, Terminal-Bench Hard 36.4%. It is particularly used for code generation.
How fast is Gemma 4 31B?
It generates about 35.2 tokens per second, with a median 1s delay before the first token. Measured independently by Artificial Analysis.
Can I self-host Gemma 4 31B?
Yes. Gemma 4 31B has open weights, so you can download it and run it on your own GPU or server with tools like Ollama, vLLM or llama.cpp — with no per-token cost.