Google DeepMind

Gemma 4 31B IT

Generally availableOpen weights262K contextReasoning

Summary

Gemma 4 31B IT is an open-weight language model from Google DeepMind, released on 2 Apr 2026. It accepts text and images and generates text, with a 262K-token context window and up to 33K output tokens per response. As of 18 Nov 2026, the lowest listed API price is $0.14 per million input tokens and $0.40 per million output tokens (Amazon Bedrock), across 10 providers we track.

Largest Gemma 4 instruction model for open, self-hosted chat and reasoning

Last verified 18 Nov 2026. Every figure on this page links to its source.

Specifications

DeveloperGoogle DeepMind
Release date2 Apr 2026 [models.dev]
StatusGenerally available [models.dev]
WeightsOpen weights [Hugging Face]
Licenseapache-2.0 [Hugging Face]
Parameters31.3B [Hugging Face]
Context window262K (262,144 tokens) [models.dev]
Max output33K tokens [models.dev]
Knowledge cutoffJan 2025 [models.dev]
InputText and images [models.dev]
OutputText [models.dev]
Reasoning modeYes [models.dev]
Tool callingYes [models.dev]
Structured outputYes [models.dev]
Hugging Facegoogle/gemma-4-31B-it [Hugging Face]

Gemma 4 31B IT API pricing by provider

USD per million tokens. Sorted by input price.

ProviderInputOutputCached inputContextSource
Gemini API
gemma-4-31b-it
FreeFree–262Kmodels.dev, 18 Nov 2026
NVIDIA NIM
google/gemma-4-31b-it
FreeFree–256Kmodels.dev, 18 Nov 2026
Amazon Bedrock
google.gemma-4-31b
$0.14$0.40–262Kmodels.dev, 18 Nov 2026
Novita AI
google/gemma-4-31b-it
$0.14$0.40–262Kmodels.dev, 18 Nov 2026
FriendliAI
google/gemma-4-31B-it
$0.14$0.40–262Kmodels.dev, 18 Nov 2026
Hugging Face Inference Providers
google/gemma-4-31B-it
$0.14$0.40–262Kmodels.dev, 18 Nov 2026
Vercel AI Gateway
google/gemma-4-31b-it
$0.14$0.40–262Kmodels.dev, 18 Nov 2026
DeepInfra
google/gemma-4-31B-it
$0.20$0.40–262Kmodels.dev, 18 Nov 2026
OpenRouter
google/gemma-4-31b-it
$0.20$0.40–262KOpenRouter, 18 Nov 2026
SiliconFlow
google/gemma-4-31B-it
$0.75$1$0.25262Kmodels.dev, 18 Nov 2026

Running Gemma 4 31B IT locally

Estimated memory, from the parameter count and the architecture in config.json on Hugging Face. Includes about 0.8 GB of runtime overhead.

QuantizationWeightsTotal, 8K contextTotal, 32K context
Q4_K_M (4-bit)19 GB22 GB26 GB
Q8_0 (8-bit)33 GB36 GB40 GB
FP16 / BF16 (16-bit)63 GB66 GB70 GB

At 4-bit with an 8K conversation, Gemma 4 31B IT needs about 22 GB, so it fits on a graphics card with 24 GB or more, such as the NVIDIA GeForce RTX 3090 24 GB. Check your own hardware, or see Best GPU for local LLMs on our sister site AIPCs.

Frequently asked questions

What is the context window of Gemma 4 31B IT?

Gemma 4 31B IT has a context window of 262,144 tokens (262K) and can generate up to 32,768 tokens in a single response, according to models.dev.

How much does the Gemma 4 31B IT API cost?

Through Gemini API, Gemma 4 31B IT costs Free per million input tokens and Free per million output tokens. Prices last verified 18 Nov 2026.

Is Gemma 4 31B IT open source?

Gemma 4 31B IT is an open-weight model: its weights are published on Hugging Face as google/gemma-4-31B-it under the apache-2.0 license. Check the license terms before commercial use.

When was Gemma 4 31B IT released?

Google DeepMind released Gemma 4 31B IT on 2 Apr 2026, according to models.dev.

Which providers offer Gemma 4 31B IT?

We track Gemma 4 31B IT on 10 providers: Gemini API, NVIDIA NIM, Amazon Bedrock, Novita AI, FriendliAI, Hugging Face Inference Providers, Vercel AI Gateway, DeepInfra, OpenRouter, SiliconFlow.

How much VRAM does Gemma 4 31B IT need?

About 22 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 36 GB at 8-bit. The 4-bit model fits on a graphics card with 24 GB or more, such as the NVIDIA GeForce RTX 3090 24 GB. These are estimates from the model's published parameter count and architecture.

Similar models from other labs

Comparable on output type, weights, release date, price and context window. Prices are the lowest listed, USD per million tokens.

ModelLabReleasedContextInput $/1MOutput $/1MWeights
NVIDIA Nemotron 3 Super 120B A12BNVIDIA11 Mar 2026262K$0.07$0.40Open
Mistral Small 4Mistral AI16 Mar 2026262K$0.15$0.60Open
Hy3 previewTencent (Hunyuan)20 Apr 2026262K$0.18$0.60Open
Jamba MiniAI21 Labs8 Jan 2026256K$0.20$0.40Open
Step 3.5 FlashStepFun29 Jan 2026262K$0.10$0.30Open
Qwen3.6 35B-A3BAlibaba (Qwen)17 Apr 2026262K$0.10$0.90Open

Change history

  • Gemini API price changed: input – → Free, output – → Free per 1M tokens
  • Knowledge cutoff changed from Dec 2025 to Jan 2025
  • Knowledge cutoff changed from Jan 2025 to Dec 2025
  • SiliconFlow price changed: input $0.13 → $0.75, output $0.40 → $1 per 1M tokens
  • OpenRouter price changed: input $0.09 → $0.20, output $0.34 → $0.40 per 1M tokens
  • DeepInfra price changed: input $0.15 → $0.20, output $0.40 → $0.40 per 1M tokens

Sources