Google Releases Gemini 4 Argon API With 1M Token Output Limit
Google has made Gemini 4 Argon generally available on the Gemini API, AI Studio and the Gemini app. The sparse MoE model writes up to 1M output tokens and leads the Vals Index at 68.90%. The post Google Releases Gemini 4 Argon API With 1M Token Output Limit appeared first on MarkTechPost.
Google has released Gemini 4 Argon, a new sparse mixture-of-experts (MoE) model with a 1M token output limit, available via the Gemini API, AI Studio, and the Gemini app. The model is designed for complex, multi-step tasks across domains like coding, finance, and cybersecurity.
What happened
Google has made Gemini 4 Argon generally available, a sparse MoE model with a 1M token output limit. It is designed for long, multi-step tasks in domains like coding, finance, and cybersecurity. The model is available via the Gemini API, AI Studio, and the Gemini app, with introductory pricing of $2 input and $10 output per million tokens.
Why it matters
Gemini 4 Argon's 1M token output limit and performance on benchmarks like the Vals Index make it a significant advancement for developers and enterprises working on complex tasks. Its availability via the Gemini API and AI Studio makes it accessible for a wide range of applications.
- Gemini 4 Argon is a sparse MoE model with a 1M token output limit.
- It leads the Vals Index at 68.90% and performs well on benchmarks like DeepSWE v1.1.
- Introductory pricing is $2 input and $10 output per million tokens.
- Argon trails on some benchmarks like FrontierSWE v2 and Terminal-Bench 4.0.
- The model is available via the Gemini API, AI Studio, and the Gemini app.
The bigger picture
Gemini 4 Argon's performance on benchmarks and its 1M token output limit can significantly impact developers and enterprises working on complex, multi-step tasks. Its availability via the Gemini API and AI Studio makes it accessible for a wide range of applications.
Gemini 4 Argon is a significant advancement in AI, offering a 1M token output limit and strong performance on key benchmarks. Its availability via the Gemini API and AI Studio makes it a valuable tool for developers and enterprises working on complex tasks.
Technical details
- Sparse MoE transformer architecture with native multimodal support.
- Trained on TPUs using JAX and ML Pathways.
- Knowledge cutoff is March 2026.
- Combined input and output token window of up to 1M tokens.
- Supports text, images, audio, and video as input, with text as output.
FAQ
What is Gemini 4 Argon?
Gemini 4 Argon is a sparse mixture-of-experts (MoE) model from Google, designed for long, multi-step tasks in domains like coding, finance, and cybersecurity.
What is the output token limit for Gemini 4 Argon?
Gemini 4 Argon can write up to 1M tokens in one response.
How much does Gemini 4 Argon cost?
Introductory pricing is $2 input and $10 output per million tokens, with prices increasing after November 30, 2026.
Where can I use Gemini 4 Argon?
Gemini 4 Argon is available via the Gemini API, AI Studio, and the Gemini app.
How does Gemini 4 Argon perform on benchmarks?
Gemini 4 Argon leads the Vals Index at 68.90% and performs well on benchmarks like DeepSWE v1.1, but trails on some benchmarks like FrontierSWE v2 and Terminal-Bench 4.0.
This summary and analysis were generated with AI from the original article and may contain errors. Read the original for the full story.
Read the full story
Originally published by MarkTechPost on Oct 14, 2026.
Comments
Sign in to join the discussion.