Automated due diligence
Run broad, multi-source investigations in the background and return a cited report to analysts or downstream systems.
Independent tool overview
Deep Research Max at a glance
Deep Research Max is Google's most comprehensive Gemini API research agent, designed for long-running background work that searches, reads, reasons and produces cited reports.
Visit the official Deep Research Max site ↗
- Product type
- Autonomous research agent API
- Model
- Gemini 3.1 Pro
- API status
- Public preview
- Typical runtime
- Minutes in background execution
- Estimated Max cost
- About $3–$7 per task
Overview
What Deep Research Max is
Deep Research Max is the higher-depth version of Google's Gemini Deep Research agent for developers. It uses Gemini 3.1 Pro and extended test-time compute to iteratively search, read, reason and refine a final report.
The product is built for asynchronous workflows such as competitive intelligence, due diligence and literature review rather than low-latency chat. It can combine the public web with files, File Search and remote MCP servers, then generate cited analysis with charts or infographics.
Deep Research Max is available in public preview through the paid Gemini API and Interactions API. It is distinct from the consumer Deep Research experience inside the Gemini app, and its cost varies with the agent's autonomous token and tool usage.
Use cases
Who Deep Research Max is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
Competitive and market intelligence
Combine web research, proprietary files and specialist data sources into a structured landscape analysis.
Research-powered applications
Embed long-horizon research into products that can tolerate multi-minute execution and variable per-task cost.
Capabilities
Core Deep Research Max features
Maximum-depth research
Uses additional compute to search more extensively and iteratively refine synthesis for complex questions.
Web and private-data grounding
Can use Google Search, URL Context, uploaded documents, File Search stores and connected data sources.
Remote MCP support
Connects to remote Model Context Protocol servers for specialized tools and proprietary data.
Native visualizations
Can generate charts, graphs and infographics alongside written findings when visualization is enabled.
Collaborative planning
Lets an application request, refine and approve a research plan before the full investigation begins.
Streaming progress
Supports streamed thought summaries, text and image deltas while the background task runs.
Process
How the Deep Research Max workflow works
Step 1
Choose the Max agent
Start an Interactions API request with the deep-research-max-preview-04-2026 agent and background execution enabled.
Step 2
Define scope and sources
Provide a precise research prompt, documents and any allowed tools such as Search, File Search or remote MCP servers.
Step 3
Review the plan when needed
Enable collaborative planning for high-stakes work, then refine or approve the proposed investigation before execution.
Step 4
Monitor the background task
Poll the interaction or stream progress, accounting for reconnection because long research jobs can outlast a single connection.
Step 5
Verify and use the report
Review citations, assumptions and visualized data before passing the result into a business or regulated workflow.
Cost
Deep Research Max pricing and free plan
Deep Research Max uses pay-as-you-go Gemini API billing. Google estimates a complex Max task at roughly $3–$7 in preview, but actual cost varies with searching, reading, model tokens and tool use.
Deep Research
Estimated $1–$3 per task
The faster agent for moderate analysis and interactive research surfaces.
- Pay as you go
- Roughly 80 searches in Google's example
- Estimated 250K input and 60K output tokens
Deep Research Max
Estimated $3–$7 per task
The comprehensive agent for extensive analysis and due diligence.
- Pay as you go
- Up to roughly 160 searches in Google's example
- Estimated 900K input and 80K output tokens
Pricing checked . Check current pricing at the source ↗
Assessment
Deep Research Max strengths and limitations
Where it stands out
- Combines public web research with files and specialized private data sources.
- Produces citations and can create native visual explanations.
- Designed for automated, asynchronous research pipelines rather than manual copy-and-paste workflows.
- Collaborative planning gives developers control before an expensive or sensitive run begins.
- Supports follow-up interactions without restarting the entire investigation.
What to consider
- The API agent is still in preview, so models, behavior and pricing can change.
- Max jobs are slower and more expensive than ordinary model calls or the standard Deep Research agent.
- Structured outputs and custom function-calling tools are not supported; remote MCP is the supported extension path.
- Research time is capped at 60 minutes, and background execution requires stored interactions.
- Web pages, uploaded files and connected tools can introduce prompt-injection or data-exposure risks, so sources and permissions need careful review.
Compare
Deep Research Max alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Consumer
Gemini Deep Research Agent
Google's faster Deep Research configuration is a better fit when latency and cost matter more than maximum coverage.
Explore Gemini Deep Research Agent →Data Analysis
Perplexity
A consumer-friendly research interface for cited answers and reports without building an API workflow.
Explore Perplexity →Business Operations
ChatGPT
A broader assistant with deep-research capabilities for users who want an interactive product rather than a Gemini API integration.
Explore ChatGPT →Questions
Deep Research Max FAQs
What is Deep Research Max?
Deep Research Max is Google's most comprehensive Gemini API research agent. It autonomously plans, searches, reads and synthesizes information into a cited report.
How is Deep Research Max different from Deep Research?
The standard Deep Research agent prioritizes speed and efficiency. Max uses more test-time compute and searches more extensively for maximum comprehensiveness.
How much does Deep Research Max cost?
Google estimates roughly $3–$7 for an extensive Max task during preview. Billing is pay as you go, so the actual amount depends on tokens, searches and other tools used.
Can Deep Research Max search private data?
Yes. The agent can use uploaded documents, File Search stores and remote MCP servers in addition to web search, subject to the application's permissions and configuration.
Is Deep Research Max available in the Gemini app?
This listing covers the developer agent exposed through the Gemini Interactions API. It is separate from the consumer Deep Research feature in the Gemini app.
Does Deep Research Max return structured JSON?
Not currently. Google's preview documentation says structured outputs are not supported.
Bottom line
Our Deep Research Max verdict
Deep Research Max is a strong fit for developers building high-depth research pipelines where coverage and synthesis matter more than latency. Its MCP, file and visualization support are unusually broad, but preview status, variable costs and agent-security risks make controlled deployment essential.
Visit Deep Research Max website ↗Similar Consumer Tools
View all Consumer AI tools →
Kimi K2.6
⚙️ Kimi K2.6 - Moonshot AI's powerful open-source coding and agent model
GPT 5.5
GPT-5.5 - OpenAI's new frontier flagship model for agentic coding and computer use
Claude Opus 4.7
Claude Opus 4.7 - Anthropic’s new top model with advanced agentic coding
Nemotron 3 Nano Omni
Nemotron 3 Nano Omni - NVIDIA's new open model combining vision, audio, and text
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.