Gemini 4 Argon
gemini-4-argon
Total Context
1M
Max Output
65.5K
Released
N/A
Input$4 / M tokens
Output$20 / M tokens
Gemma 4 31B
gemma-4-31b-itGemma 4 26B A4B is an efficient Mixture-of-Experts (MoE) model from Google DeepMind, combining 26B-class model capacity with approximately 4B active parameters during inference. It delivers strong reasoning, coding, multimodal understanding, and agentic capabilities while offering faster and more efficient inference than larger dense models. With support for text and image inputs, a 256K context window, native function calling, and more than 140 languages, it is well suited for AI agents, coding applications, long-context processing, and high-throughput workloads.
Context Window
262.1K tokens
Maximum Output
32.8K tokens
Release Date
Apr 2, 2026
Modalities
Gemma 4 31B Pricing
| Input Price | Output Price | Cache Read |
|---|---|---|
| $0.75/M | $1/M | $0.75/M |
Gemma 4 31B API Capabilities
Reasoning
Supported
Tool calling
Supported
Temperature parameter
Supported
Attachments
Supported
Knowledge Base
—
Endpoint Protocols
Completions APIResponses APIMessages API
Model Highlights
Dense reasoning, long context, and visual understanding.
Dense Reasoning
A 30.7B-parameter dense model with configurable thinking, designed for reasoning and coding.
Long Context
Supports a 256K-token context window with interleaved local and global attention.
Visual Understanding
Combines text and image inputs to generate text, supporting visual questions and document interpretation.
Use Cases
Practical workloads for coding, documents, and images.
Coding Assistance
Draft code, explain implementations, and propose fixes using supplied source files and requirements.
Document Synthesis
Summarize lengthy materials, compare sections, and answer questions using documents supplied in context.
Visual Document Review
Interpret screenshots, charts, and document images to produce descriptions and answers for human review.
How to Use Gemma 4 31B via the TokenHub API
import OpenAI from "openai"
const client = new OpenAI({
apiKey: process.env.TOKENHUB_API_KEY,
baseURL: "https://us-api.tokenhub.com/v1",
})
const result = await client.chat.completions.create({})
console.log(result.choices[0]?.message?.content)FAQs
Model capabilities, tradeoffs, and TokenHub usage.
What is Gemma 4 31B?+
It is Google DeepMind’s instruction-tuned, open-weight dense model with 30.7B parameters. It accepts text and images and generates text; its weights use the Apache 2.0 license.
Which workloads suit its strengths?+
Coding assistance, multi-step reasoning, long-document analysis, and visual questions are suitable starting points. Configurable thinking and a 256K context window support these workflows.
What limitations should I consider?+
Outputs can be inaccurate and require review. This variant has no native audio input. Long contexts and thinking can increase resource use and latency; hosted limits may be lower than the model’s maximum.
How can I use it on TokenHub, and what does it cost?+
Confirm availability, the exact model ID, supported inputs, and pricing in your TokenHub account. If listed, use the endpoint and credentials shown in its integration guide. Open weights do not imply free hosted inference.
How should I choose between the dense and MoE variants?+
Evaluate the dense variant when reasoning and coding quality are priorities. Compare it with the 26B A4B MoE variant when inference efficiency matters, using the same prompts and deployment settings.
Ready to use Gemma 4 31B?
Use one API key to access Gemma 4 31B and more AI models through TokenHub.
Media and Demos
Official posts and community demonstrations. Community observations reflect individual setups.
X (Twitter)
Google AI Studio ·
https://t.co/2CJ04Qlfdx
View originalPrince Canuma ·
Gemma 4 31B running with TurboQuant KV cache on MLX 🔥 128K context: → KV Memory: 13.3 GB → 4.9 GB (63% reduction) → Peak Memory: 75.2 GB → 65.8 GB (-9.4 GB) → Quality preserved TurboQuant compression scales with sequence length, so the longer the context, the bigger the
View originalGoogle Gemma ·
Gemma 4 is live on @cerebras, the fastest multimodal inference ever! Running on Gemma 4 31B open-weight model at a blistering 1,500+ tokens/sec. That's a 15x speedup, unlocking real-time visual and agentic loops without the GPU lag. https://t.co/jtMkmPbbsG
https://pbs.twimg.com/amplify_video_thumb/2076713295294373888/img/ag8ksefiL-46gnOj.jpgView originalReddit
Is Gemma 4 31b overkill for a personal assistant/RAG?
vick2djax ·
Read full post
My main project is an all in one chatbot that focuses on research with a huge RAG and web browsing abilities (I ingest all my books, most of Wiki, all the big data sets for research papers and such). The idea is to have an auditable process that helps me sort through misinformation and basically the garbage that most of the internet has become. I’ve got RSS feeds feeding the RAG daily from places like Reuters to help get past any model’s knowledge cutoff and just have good information come in. But I don’t want it to be a clipping service that just cuts sentences out of pages and glues them together on a new one to make a scrapbook of paragraphs. I want it to take it all as evidence and reason through it. Be able to go back and forth with me as I drill in deeper into a topic. Additionally have it wired up to things like a big memory system to remember all my chats and piece together personal stuff about me so I can better plan things. Hooked into the documentation for my git projects. Etc. 🚨Gemma 4 26b MoE is matching and/or beating the 31b dense on every damn test I come up with🚨 This sub lead me to believe that dense good, moe bad. Moe dumb. I feel like maybe my testing is wrong or something. The responses I get out of the models take turns on giving me the best responses as well. So confusing. Anybody else with the same type of project that has run into the same thing? Do I just take the 2x/3x speed of moe and skip out on the dense? To note, this project isn’t about coding. Zero roleplaying although I do enjoy creative brainstorming on projects. Hardware: solo 3090. I’m able to fit Gemma 4 31b at Q4 cache with 65k+ ctx with QAT & MTP. From what I understand, the QAT is supposed to make the q4 KV not so much of a bad thing for Gemma. QAT UD-Q4_K_XL Note: I was originally using qwen3.6 and had the same issues of 27b getting outperformed by the 35b. I swapped to Gemma as I kept reading it was much better at writing so figured it would make query results easier to read.
Which model to use? Gemma 4 26B A4B or Gemma 4 31B?
Yaniekk ·
Read full post
Hey everyone! I'm currently working on a cooking web app where users can post their own recipes. I've already integrated the OpenAI Moderations API to filter out unwanted content (such as NSFW, violence, hate speech, etc.). It works great, but now I need a way to evaluate the overall quality of a post. Specifically, I'd like to check whether the recipe makes sense or is just gibberish, whether it's actually about cooking, whether it's complete, whether the attached image matches the recipe, and so on. I was thinking about using an LLM for this, and I found two free models on OpenRouter: Gemma 4 26B A4B and Gemma 4 31B. I'm looking for something that's accurate, reasonably fast, and suitable for this kind of validation in a production app. Which of these models would you recommend? Or is there another free model on OpenRouter that would be a better choice?
Gemma 4 31B Base or Gemma 4 31B IT, Which one is better for RP?
Public-Speed125 ·
I'm planning to use Gemma 4 31B with SillyTavern, mostly for RP, storytelling, and long conversations. I'm wondering which version people here prefer Gemma 4 31B Base or Gemma 4 31B IT and also any differences in creativity, coherence, or censorship? I'd love to hear your experiences before I commit to downloading one. Thanks!
View originalYouTube
🦞在OpenClaw实测谷歌开源大模型Gemma 4！256K上下文+多模态输入，小龙虾里实战测试多Agent协作与浏览器自动化全流程！31B参数碾压其他开源大模型！五大维度深度评分结果出人意料
AI超元域
View originalTesting Gemma 4 31B with OpenCode
NightShift Tech
View original谷歌杀疯了！Gemma 4彻底开源，31B模型挑战百亿巨兽！Google's Gone Wild! Gemma 4 Fully Open-Sourced
大雷早上好
View original