# MiroThinker
**Repository Path**: act6ps/MiroThinker
## Basic Information
- **Project Name**: MiroThinker
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No
## Statistics
- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-01-23
- **Last Updated**: 2026-01-23
## Categories & Tags
**Categories**: Uncategorized
**Tags**: None
## README
[](https://dr.miromind.ai/)
[](https://miromind.ai/#blog)
[](https://huggingface.co/datasets/miromind-ai/MiroVerse-v0.1)
[](https://github.com/MiroMindAI)
[](https://miromind.ai/)
[](https://discord.com/invite/GPqEnkzQZd)
### π [Try our Demo!](https://dr.miromind.ai/)
> **MiroThinker** is an open source deep research agent optimized for research and prediction. It achieves a 80.8% Avg@8 score on the challenging GAIA benchmark.
The project currently comprises four key components:
- π‘ **MiroThinker**: An open source deep research agent optimized for research and prediction. It achieves a 80.8% Avg@8 score on the challenging GAIA benchmark. See [Quick Start](#-quick-start).
- π€ **MiroFlow**: An agent framework that enables tool-use agent tasks, featuring a reproducible GAIA score of 82.4%. See [MiroFlow](https://github.com/MiroMindAI/MiroFlow) for details.
- π **MiroVerse**: A premium open-source training dataset with 147k samples supporting research agent training. See [MiroVerse](https://huggingface.co/datasets/miromind-ai/MiroVerse-v0.1) on HuggingFace.
## π Table of Contents
- π° [News & Updates](#-news--updates)
- π [Introduction](#-introduction)
- β¨ [Key Features](#-key-features)
- π [Performance on Benchmarks](#-performance-on-benchmarks)
- π [Quick Start](#-quick-start)
- π [Benchmark Evaluation](#-benchmark-evaluation)
- π¬ [Trace Collection](#-trace-collection)
- β [FAQ & Troubleshooting](#-faq--troubleshooting)
- π [License](#-license)
- π [Acknowledgments](#-acknowledgments)
## π° News & Updates
- **\[2026-01-05\]** ππ We release [MiroThinker-v1.5](https://huggingface.co/collections/miromind-ai/mirothinker-v15), a series of open source deep research agents optimized for financial prediction. [MiroThinker-v1.5-30B](https://huggingface.co/miromind-ai/MiroThinker-v1.5-30B) surpasses Kimi-K2-Thinking on BrowseComp-ZH at much lower cost, using only 1/30 of the parameters. [MiroThinker-v1.5-235B](https://huggingface.co/miromind-ai/MiroThinker-v1.5-235B) scores 39.2% on HLE-Text, 69.8% on BrowseComp, 71.5% on BrowseComp-ZH, and 80.8% on GAIA-Val-165, setting a new state-of-the-art among search agents.
- **\[2025-11-13\]** π [MiroThinker-v1.0](https://huggingface.co/collections/miromind-ai/mirothinker-v10) is now released! Introducing **interactive scaling** as a third dimension of performance improvement, MiroThinker v1.0 supports 256K context window and up to 600 tool calls per task. Available in 8B, 30B, and 72B parameter scales, achieving 37.7%, 47.1%, 55.6%, and 81.9% on HLE-Text, BrowseComp, BrowseComp-ZH, and GAIA-Text-103, respectively. See [Technical Report](https://arxiv.org/abs/2511.11793) for more details.
- **\[2025-09-11\]** MiroThinker-72B-Preview ranked 4th in this week's FutureX benchmark. See [FutureX](https://futurex-ai.github.io/).
π Click to expand older updates
- **\[2025-09-08\]** [MiroThinker-v0.2](https://huggingface.co/collections/miromind-ai/mirothinker-v02) is now released, achieving open-source SOTA performance across multiple benchmarks, including HLE (17.8%), HLE-Text-Only (19.1%), BrowseComp-EN (17.2%), BrowseComp-ZH (29.4%), XBench-DeepSearch (56.0%), and Frames (74.8%).
- **\[2025-09-07\]** We supported more benchmarks, including [BrowseComp-ZH](https://arxiv.org/abs/2504.19314), [XBench-DeepSearch](https://xbench.org/agi/aisearch), and [FutureX](https://futurex-ai.github.io/). We plan to add more benchmarks in the future.
- **\[2025-08-22\]** Introducing streamlined deployment options for MiroThinker with optimized resource usage and faster startup times. Experience the interactive demo: [π Try Gradio Demo](apps/gradio-demo)
- **\[2025-08-08\]** [MiroThinker-v0.1](https://huggingface.co/collections/miromind-ai/mirothinker-v01-689301b6d0563321862d44a1) released.
## π Introduction
### MiroThinker-v1.5
MiroThinker v1.5 is the world-leading open-source search agent that advances tool-augmented reasoning through **interactive scaling** β training the agent to handle deeper and more frequent agent-environment interactions as a third dimension of performance improvement, beyond model size and context length.
**Key Features**
- π MiroThinker v1.5 supports a 256K context window, long-horizon reasoning, and deep multi-step analysis.
- π§ Handles up to 400 tool calls per task β a substantial improvement over previous open-source research agents.
- π¦ Released in 30B and 235B parameter scales, accompanied by a comprehensive suite of tools and workflows to flexibly support diverse research settings and compute budgets.
| Agent Name | Base Agent | Max Context | Max Tool Calls | HF Link |
|:---------------------:|:-----------------------------:|:-----------:|:--------------:|:------------------------------------------------------------------:|
| MiroThinker-v1.5-30B | Qwen3-30B-A3B-Thinking-2507 | 256K | 400 | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-v1.5-30B) |
| MiroThinker-v1.5-235B | Qwen3-235B-A22B-Thinking-2507 | 256K | 400 | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-v1.5-235B) |
MiroThinker v1.5 demonstrates strong general-research performance across a broad range of benchmarks, achievingΒ 39.2%,Β 69.8%, 71.5%, andΒ 80.8%Β on HLE-Text, BrowseComp, BrowseComp-ZH, and GAIA-Val-165, respectively. These results surpass previous open-source agents and set the new world-leading BrowseComp performance.
### MiroThinker-v1.0
π¦ Click to expand MiroThinker-v1.0 details
Unlike previous agents that scale only model size or context length, MiroThinker v1.0 introduces **interactive scaling** at the agent level, systematically training the agent to handle deeper and more frequent agentβenvironment interactions as a third dimension of performance improvement. Interactive scaling leverages environment feedback and external information acquisition to correct errors and refine trajectories.
### β¨ Key Features
- π **256K Context Window**: Supports long-horizon reasoning and deep multi-step analysis
- π§ **600 Tool Calls**: Handles up to 600 tool calls per task β a substantial improvement over previous open-source research agents
- π¦ **Multiple Scales**: Released in 8B, 30B, and 72B parameter scales, accompanied by a comprehensive suite of tools and workflows to flexibly support diverse research settings and compute budgets
| Agent Name | Base Agent | Max Context | Max Tool Calls | HF Link |
|:--------------------:|:---------------------------:|:-----------:|:--------------:|:------------------------------------------------------------------:|
| MiroThinker-v1.0-8B | Qwen3-8B | 256K | 600 | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-v1.0-8B) |
| MiroThinker-v1.0-30B | Qwen3-30B-A3B-Thinking-2507 | 256K | 600 | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-v1.0-30B) |
| MiroThinker-v1.0-72B | Qwen2.5-72B-Instruct | 256K | 600 | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-v1.0-72B) |
MiroThinker v1.0 demonstrates strong general-research performance across a broad range of benchmarks, achieving **37.7%**, **47.1%**, **55.6%**, and **81.9%** on HLE-Text, BrowseComp, BrowseComp-ZH, and GAIA-Text-103, respectively. These results surpass previous open-source agents and narrow the gap with commercial counterparts such as **GPT-5-high**.
### MiroThinker-v0.2
π¦ Click to expand MiroThinker-v0.2 details
In this new version, we introduced three key improvements:
- π **Richer training data** from both English and Chinese sources, yielding significant gains in benchmark performance and generalization
- π― **Unified DPO training** with a single preference dataset across all agents
- π **Extended context length** from 40k to 64k for more challenging multi-turn tool-use tasks
Compared to v0.1, MiroThinker v0.2 delivers consistent gains across benchmarks. For example, scores improved from **57.3 β 64.1** on **GAIA-Text-103** and from **17.0 β 29.4** on **BrowseComp-ZH**, reflecting substantial advancements in the modelβs general research agent capabilities.
| Agent Name | Base Agent | Max Context | HF Link |
|:------------------------:|:---------------------:|:-----------:|:----------------------------------------------------------------------:|
| MiroThinker-4B-SFT-v0.2 | Qwen3-4B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-4B-SFT-v0.2) |
| MiroThinker-4B-DPO-v0.2 | Qwen3-4B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-4B-DPO-v0.2) |
| MiroThinker-8B-SFT-v0.2 | Qwen3-8B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-8B-SFT-v0.2) |
| MiroThinker-8B-DPO-v0.2 | Qwen3-8B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-8B-DPO-v0.2) |
| MiroThinker-14B-SFT-v0.2 | Qwen3-14B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-14B-SFT-v0.2) |
| MiroThinker-14B-DPO-v0.2 | Qwen3-14B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-14B-DPO-v0.2) |
| MiroThinker-32B-SFT-v0.2 | Qwen3-32B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-32B-SFT-v0.2) |
| MiroThinker-32B-DPO-v0.2 | Qwen3-32B | 64K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-32B-DPO-v0.2) |
### MiroThinker-v0.1
π¦ Click to expand MiroThinker-v0.1 details
Performance of Open-Source Agents on GAIA-Validation Benchmark.
We have released the **MiroThinker v0.1** series, including both SFT and DPO variants at parameter scales of **8B**, **14B**, and **32B**. Notably, MiroThinker v0.1 achieves **state-of-the-art performance** among open-source models on the [GAIA benchmark](https://huggingface.co/datasets/gaia-benchmark/GAIA), a rigorous evaluation suite for advanced agentic capabilities, demonstrating its strength in long-context, decision-intensive, and real-world task scenarios.
| Agent Name | Base Agent | Max Context | HF Link |
| :-----------------------: |:----------:|:-----------:| :--------------------------------------------------------------------:|
| MiroThinker-8B-SFT-v0.1 | Qwen3-8B | 40K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-8B-SFT-v0.1) |
| MiroThinker-8B-DPO-v0.1 | Qwen3-8B | 40K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-8B-DPO-v0.1) |
| MiroThinker-14B-SFT-v0.1 | Qwen3-14B | 40K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-14B-SFT-v0.1) |
| MiroThinker-14B-DPO-v0.1 | Qwen3-14B | 40K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-14B-DPO-v0.1) |
| MiroThinker-32B-SFT-v0.1 | Qwen3-32B | 40K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-32B-SFT-v0.1) |
| MiroThinker-32B-DPO-v0.1 | Qwen3-32B | 40K | [π€ link](https://huggingface.co/miromind-ai/MiroThinker-32B-DPO-v0.1) |
## β¨ Key Features
### π€ **MiroThinker-Optimized Framework**
- π **Fully Open-Source Agent Framework**: Complete transparency with open framework and open agents
- π **Tool Integration**: Seamless integration with external tools and APIs
- π **Trace Collection**: Comprehensive logging and analysis of agent interactions with elapsed time and estimated completion time displayed in minutes. Ready for SFT and DPO
- π **Benchmark Evaluation**: Extensive testing across multiple benchmark datasets
### π **Comprehensive Benchmark Suite**
π Click to expand benchmark list
- **GAIA Validation**: A benchmark for General AI Assistants. ([paper](https://arxiv.org/abs/2311.12983))
- **GAIA-Text-103**: A subset of GAIA Validation for text-only tasks. ([paper](https://arxiv.org/abs/2505.22648))
- **HLE**: Humanity's Last Exam. ([paper](https://arxiv.org/abs/2501.14249))
- **HLE-Text-2158**: A subset of HLE for text-only tasks. ([paper](https://arxiv.org/abs/2501.14249))
- **HLE-Text-500**: A subset of HLE for text-only tasks, created by [WebThinker](https://arxiv.org/pdf/2504.21776). ([paper](https://arxiv.org/pdf/2504.21776))
- **BrowseComp-EN**: Web browsing and comprehension tasks. ([paper](https://arxiv.org/abs/2504.12516))
- **BrowseComp-ZH**: A Chinese version of BrowseComp. ([paper](https://arxiv.org/abs/2504.19314))
- **WebWalkerQA**: Web navigation and question answering. ([paper](https://arxiv.org/abs/2501.07572))
- **Frames**: Factuality, Retrieval, And reasoning MEasurement Set. ([paper](https://arxiv.org/abs/2409.12941))
- **XBench-DeepSearch**: A benchmark for deep research agents. ([website](https://xbench.org/agi/aisearch))
- **FutureX**: A live benchmark designed for predicting unknown future. ([website](https://futurex-ai.github.io/))
- **SEAL-0**: A benchmark for evaluating LLMs on conflicting-evidence web questions. ([paper](https://arxiv.org/abs/2506.01062))
- **AIME2025**: American Invitational Mathematics Examination 2025. ([website](https://artificialanalysis.ai/evaluations/aime-2025))
- **DeepSearchQA**: Google's Deep Search Question Answering benchmark. ([paper](https://arxiv.org/abs/2505.20827))
## π Performance on Benchmarks
### MiroThinker-v1.5
> To prevent potential information leakage (e.g., searching benchmark answers from HuggingFace), access to HuggingFace has been explicitly disabled in these tools.
> We further perform canary string testing on the tool outputs of all trajectories and disregard any trajectory found to be contaminated, treating it as an incorrect answer.
### MiroThinker-v1.0
π¦ Click to expand MiroThinker-v1.0 details
### MiroThinker-v0.2
π¦ Click to expand MiroThinker-v0.2 details
#### Comparison with SOTA Research Agents
#### GAIA Benchmark
### MiroThinker-v0.1
π¦ Click to expand MiroThinker-v0.1 details
#### GAIA Benchmark
| **Method** | Text-103
Best Pass@1 | Text-103
Pass@1 (Avg@8) | Val-165
Best Pass@1 | Val-165
Pass@1 (Avg@8) |
|------------------------------|:-----------------------:|:--------------------------:|:----------------------:|:-------------------------:|
| **πΉββ 7B/8B Agents ββ** | | | | |
| Search-o1-7B | 17.5 | - | - | - |
| R1-Searcher-7B | 20.4 | - | - | - |
| WebDancer-7B | 31.0 | - | - | - |
| WebSailor-7B | 37.9 | - | - | - |
| CK-Pro-8B | 40.3 | - | 32.7 | - |
| **MiroThinker-8B-SFT-v0.1** | 44.7 | 40.1 | 34.6 | 31.8 |
| + Commercial Tools | 46.6 | 42.1 | 37.6 | 33.9 |
| **MiroThinker-8B-DPO-v0.1** | 46.6 | 44.8 | 37.0 | 35.4 |
| + Commercial Tools | **50.5** | **46.7** | **38.2** | **35.9** |
| **πΉββ 14B Agents ββ** | | | | |
| **MiroThinker-14B-SFT-v0.1** | 47.6 | 44.4 | 37.0 | 34.4 |
| + Commercial Tools | 49.5 | 47.5 | 41.8 | 39.8 |
| **MiroThinker-14B-DPO-v0.1** | 48.5 | 46.6 | 42.4 | 39.2 |
| + Commercial Tools | **52.4** | **48.5** | **45.5** | **42.0** |
| **πΉββ 32B Agents ββ** | | | | |
| Qwen3-32B | 31.1 | 26.7 | 29.7 | 26.4 |
| Search-o1-32B | 28.2 | - | - | - |
| WebThinker-32B-RL | 48.5 | - | - | - |
| WebDancer-QwQ-32B | 51.5 | - | - | - |
| WebSailor-32B | 53.2 | - | - | - |
| WebShaper-QwQ-32B | 53.3 | - | - | - |
| **MiroThinker-32B-SFT-v0.1** | 55.3 | 51.3 | 44.9 | 42.7 |
| + Commercial Tools | 58.3 | 54.2 | 48.5 | 45.8 |
| **MiroThinker-32B-DPO-v0.1** | 57.3 | 54.1 | 48.5 | 45.9 |
| + Commercial Tools | **60.2** | **57.9** | **50.9** | **48.9** |
1. Following the practices of WebThinker, WebAgents, and CognitiveKernel, we report the Best Pass@1, the highest score across three runs, which often reflects stronger performance, though it may exhibit some variability. To provide a more stable measure, we additionally report Pass@1 (Avg@8), which offers greater consistency at the cost of slightly lower scores.
1. For consistency with prior open-source works, we evaluate GAIA-Text-103 using the WebAgents LLM-as-a-Judge template, and report results on GAIA-Val-165 using the official GAIA scorer script.
1. By default, we use open-source tools wherever possible, except for the code tool [E2B](https://github.com/e2b-dev/E2B) and the Google search tool [Serper](https://serper.dev/). We use [Whisper](https://huggingface.co/openai/whisper-large-v3-turbo), [Qwen2.5-VL-72B-Instruct](https://huggingface.co/Qwen/Qwen2.5-VL-72B-Instruct), and [Qwen3-235B-A22B-Thinking-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507) in our implementation. The framework can be easily extended to other open-source tools of your choice.
1. Replacing these open-source tools with commercial alternatives can yield performance gains. Commercial tools were mainly used for multimodal capabilities and certain complex reasoning subtasks. The majority of tasks, including planning, browsing, refinement, navigation, and more, were handled by our agents.
#### More Benchmarks
| Method | HLE
Pass@1 | Frames
Pass@1 | BrowseComp
Pass@1 | BrowseComp-ZH
Pass@1 | WebWalkerQA
Pass@1 |
|------------------------------|:-------------:|:----------------:|:--------------------:|:-----------------------:|:---------------------:|
| OpenAI Deep Research | 26.6 | - | 51.5 | 42.9 | - |
| Gemini Deep Research | 26.9 | - | - | - | - |
| Kimi-Researcher | 26.9 | 78.8 | - | - | - |
| | | | | | |
| WebDancer-7B | - | - | - | - | 36.0 |
| WebSailor-7B | - | - | 6.7 | 14.2 | - |
| **MiroThinker-8B-SFT-v0.1** | - | 58.0 | 5.5 | 9.3 | 41.3 |
| **MiroThinker-8B-DPO-v0.1** | - | 64.4 | 8.7 | 13.6 | 45.7 |
| | | | | | |
| WebThinker-32B-RL | - | - | - | - | 46.5 |
| WebDancer-QwQ-32B | - | - | 3.8 | 18.0 | 47.9 |
| WebSailor-32B | - | - | 10.5 | 25.5 | - |
| WebShaper-32B | - | - | - | - | 51.4 |
| **MiroThinker-32B-SFT-v0.1** | 10.2 | 70.4 | 10.6 | 13.8 | 45.7 |
| **MiroThinker-32B-DPO-v0.1** | 11.8 | 71.7 | 13.0 | 17.0 | 49.3 |
1. MiroThinkerβs performance was tested with this repository and open-source tools; other agentsβ results are from their papers and official sites.
1. As [MiroVerse-v0.1](https://huggingface.co/datasets/miromind-ai/MiroVerse-v0.1) mainly contains English data, the agentβs Chinese capability is limited. We plan to add more Chinese data to improve performance in the next version.
## π Quick Start
### Prerequisites
- π **Python 3.10+**
- π¦ **uv package manager** ([Installation guide](https://github.com/astral-sh/uv))
- π **Required API keys** (see configuration section below)
### Installation
```bash
# Clone the repository
git clone https://github.com/MiroMindAI/MiroThinker
cd MiroThinker
# Setup environment
cd apps/miroflow-agent
uv sync
# Configure API keys
cp .env.example .env
# Edit .env with your API keys (SERPER_API_KEY, JINA_API_KEY, E2B_API_KEY, etc.)
```
> **π Environment Variables**: See [Tool Configuration](#tool-configuration) section for required API keys.
### Tool Configuration
#### Minimal Configuration for MiroThinker v1.5 and v1.0
| Server | Description | Tools Provided | Required Environment Variables |
|:-------|:------------|:---------------|:-------------------------------|
| **`tool-python`** | Execution environment and file management (E2B sandbox) | `create_sandbox`, `run_command`, `run_python_code`, `upload_file_from_local_to_sandbox`, `download_file_from_sandbox_to_local`, `download_file_from_internet_to_sandbox` | `E2B_API_KEY` |
| **`search_and_scrape_webpage`** | Google search via Serper API | `google_search` | `SERPER_API_KEY`, `SERPER_BASE_URL` |
| **`jina_scrape_llm_summary`** | Web scraping with LLM-based information extraction | `scrape_and_extract_info` | `JINA_API_KEY`, `JINA_BASE_URL`, `SUMMARY_LLM_BASE_URL`, `SUMMARY_LLM_MODEL_NAME`, `SUMMARY_LLM_API_KEY` |
**Minimal `.env` configuration example:**
```bash
# Required for MiroThinker v1.5 and v1.0 (minimal setup)
SERPER_API_KEY=your_serper_key
SERPER_BASE_URL="https://google.serper.dev"
JINA_API_KEY=your_jina_key
JINA_BASE_URL="https://r.jina.ai"
E2B_API_KEY=your_e2b_key
# Required for jina_scrape_llm_summary
# Note: Summary LLM can be a small model (e.g., Qwen3-14B or GPT-5-Nano)
# The choice has minimal impact on performance, use what's most convenient
SUMMARY_LLM_BASE_URL="https://your_summary_llm_base_url/v1/chat/completions"
SUMMARY_LLM_MODEL_NAME=your_llm_model_name # e.g., "Qwen/Qwen3-14B" or "gpt-5-nano"
SUMMARY_LLM_API_KEY=your_llm_api_key # Optional, depends on LLM provider
# Required for benchmark evaluation (LLM-as-a-Judge)
OPENAI_API_KEY=your_openai_key # Required for running benchmark evaluations
OPENAI_BASE_URL="https://api.openai.com/v1" # Optional, defaults to OpenAI's API
```
> **π‘ Why this is minimal**: These 3 MCP servers cover the core capabilities needed for research tasks: web search, content extraction, and code execution. All other servers are optional enhancements.
>
> **π€ Summary LLM**: The `SUMMARY_LLM` can be a small model like Qwen3-14B or GPT-5-Nano. The choice has minimal impact on overall performance, use whichever is most convenient for your setup.
>
> **π For Benchmark Evaluation**: If you plan to run benchmark evaluations, you also need `OPENAI_API_KEY` (and optionally `OPENAI_BASE_URL`) for LLM-as-a-Judge functionality used in evaluation scripts.
>
> **πΌοΈ For GAIA Multimodal Tasks**: GAIA-Val-165 includes tasks with image/audio/video files. Since MiroThinker is a text-only LLM, GPT-4o is used to pre-process these files into text descriptions. The same `OPENAI_API_KEY` is used for both this preprocessing and LLM-as-a-Judge.
>
> **π For more details**: See [MiroFlow Tools README](libs/miroflow-tools/README.md) for complete documentation of all available tools.
π§ Click to expand additional available tools
The following optional tools are available but were not used in MiroThinker v1.5 and v1.0 evaluation:
| Server Name | Type | Description |
|:---------------------|:-------------|:--------------------------------------------|
| `tool-vqa` | Commercial | Vision processing using Claude |
| `tool-vqa-os` | Open-Source | Vision processing (open-source alternative) |
| `tool-transcribe` | Commercial | Audio transcription using OpenAI |
| `tool-transcribe-os` | Open-Source | Audio transcription using Whisper |
| `tool-reasoning` | Commercial | Reasoning engine using Claude |
| `tool-reasoning-os` | Open-Source | Reasoning engine (open-source alternative) |
| `tool-reading` | Open-Source | Document reading using MarkItDown |
| `tool-google-search` | Commercial | Web search using Google + scraping |
| `tool-sogou-search` | Commercial | Web search using Sogou (Chinese) |
> **π Local Deployment**: For instructions on deploying open-source tools (`tool-vqa-os`, `tool-transcribe-os`, `tool-reasoning-os`) locally, see [Local Tool Deployment Guide](assets/LOCAL-TOOL-DEPLOYMENT.md).
See the [MiroFlow Tools README](libs/miroflow-tools/README.md) for complete documentation of all available tools.
#### Pre-configured Agent Settings
The `apps/miroflow-agent/conf/agent/` directory contains several pre-configured agent settings. Each configuration uses different tools and requires corresponding environment variables in your `.env` file.
> **π‘ Recommended**: For MiroThinker v1.5, use `mirothinker_v1.5_keep5_max200` (with context management, recommended for most tasks) or `mirothinker_v1.5_keep5_max400` (only used for BrowseComp and BrowseComp-ZH). For v1.0, use `mirothinker_v1.0_keep5` (with context management). All use minimal configuration with only 3 MCP servers.
| Configuration | Description | Max Turns | Context Retention | Required Environment Variables | Recommended For |
|:---------------------------------------|:------------|:----------|:------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------|:----------------|
| **`mirothinker_v1.5_keep5_max200`** β | Single-agent with context management | 200 | Keep 5 most recent | `SERPER_API_KEY`, `SERPER_BASE_URL`, `JINA_API_KEY`, `JINA_BASE_URL`, `E2B_API_KEY`, `SUMMARY_LLM_BASE_URL`, `SUMMARY_LLM_MODEL_NAME`, `SUMMARY_LLM_API_KEY` | **v1.5 (recommended for most tasks)** |
| **`mirothinker_v1.5_keep5_max400`** β | Single-agent with context management | 400 | Keep 5 most recent | Same as above | **v1.5 (for BrowseComp & BrowseComp-ZH)** |
| **`mirothinker_v1.5`** | Single-agent for MiroThinker v1.5 | 600 | Keep all results | Same as above | **v1.5** |
| **`mirothinker_v1.0_keep5`** | Single-agent with context management | 600 | Keep 5 most recent | Same as above | **v1.0** |
| **`mirothinker_v1.0`** | Single-agent for MiroThinker v1.0 | 600 | Keep all results | Same as above | **v1.0** |
π¦ Click to expand legacy configurations (v0.1/v0.2)
| Configuration | Description | Max Turns | Context Retention | Required Environment Variables | Recommended For |
|:-------------------------|:------------|:----------|:------------------|:-------------------------------|:----------------|
| **`multi_agent`** | Multi-agent with commercial tools (v0.1/v0.2) | 50 | Keep all results | `E2B_API_KEY`, `ANTHROPIC_API_KEY`, `ANTHROPIC_BASE_URL`, `OPENAI_API_KEY`, `OPENAI_BASE_URL`, `SERPER_API_KEY`, `SERPER_BASE_URL`, `JINA_API_KEY`, `JINA_BASE_URL` | v0.1/v0.2 |
| **`multi_agent_os`** | Multi-agent with open-source tools (v0.1/v0.2) | 50 | Keep all results | `E2B_API_KEY`, `VISION_API_KEY`, `VISION_BASE_URL`, `VISION_MODEL_NAME`, `WHISPER_API_KEY`, `WHISPER_BASE_URL`, `WHISPER_MODEL_NAME`, `REASONING_API_KEY`, `REASONING_BASE_URL`, `REASONING_MODEL_NAME`, `SERPER_API_KEY`, `SERPER_BASE_URL`, `JINA_API_KEY`, `JINA_BASE_URL` | v0.1/v0.2 |
> **π‘ Note**: All environment variables are listed in `apps/miroflow-agent/.env.example`. Copy it to `.env` and fill in the values for the tools you plan to use.
#### Creating Custom Tool Configurations
π§ Click to expand custom tool configuration guide
You can create your own YAML configuration file to freely combine MCP servers. Here's how:
1. **Create a new YAML file** in `apps/miroflow-agent/conf/agent/`:
```yaml
# conf/agent/my_custom_config.yaml
defaults:
- default
- _self_
main_agent:
tools:
- tool-python # Execution environment
- search_and_scrape_webpage # Google search
- jina_scrape_llm_summary # Web scraping with LLM
- tool-vqa # Vision processing (optional)
- tool-transcribe # Audio processing (optional)
- tool-reasoning # Reasoning engine (optional)
- tool-reading # Document reading (optional)
max_turns: 400 # Maximum number of turns
sub_agents:
agent-browsing: # Optional sub-agent
tools:
- tool-google-search
- tool-vqa
- tool-reading
- tool-python
max_turns: 50
keep_tool_result: -1 # Context retention budget: -1 keeps all tool results, or specify K to keep only the K most recent tool responses
```
> **π‘ Context Retention Strategy**: The `keep_tool_result` parameter implements a **recency-based context retention** strategy. In the standard ReAct paradigm, all tool outputs are retained in the message history, which can lead to inefficient context utilization. Empirically, we observe that the agent's subsequent actions depend primarily on recent observations rather than distant ones. This strategy retains only the most recent K tool responses (where K is the `keep_tool_result` value) while preserving the complete sequence of thoughts and actions.
>
> **Benefits:**
>
> - β
Preserves the reasoning and action trace
> - β
Focuses the agent's attention on the most contextually relevant observations
> - β
Frees additional context space for extended reasoning and deeper tool-use trajectories
> - β
Does not lead to performance degradation while allowing more context space for interactive scaling
>
> **Usage:** Set `keep_tool_result: -1` to keep all tool results, or specify a positive integer K (e.g., `keep_tool_result: 5`) to keep only the K most recent tool responses.
2. **Use your custom configuration** when running evaluations:
```bash
cd apps/miroflow-agent
uv run main.py llm=qwen-3 agent=my_custom_config llm.base_url=https://your_base_url/v1
```
3. **Configure environment variables** in `.env` based on the tools you use.
All available environment variables are listed in `apps/miroflow-agent/.env.example`. Copy it to `.env` and configure the variables according to your chosen configuration:
```bash
cd apps/miroflow-agent
cp .env.example .env
# Edit .env with your actual API keys
```
**For MiroThinker v1.5** (`mirothinker_v1.5_keep5_max200.yaml`, `mirothinker_v1.5_keep5_max400.yaml`, or `mirothinker_v1.5.yaml`) and **v1.0** (`mirothinker_v1.0_keep5.yaml` or `mirothinker_v1.0.yaml`), see the [Minimal Configuration](#minimal-configuration-for-mirothinker-v15-and-v10) section above for the complete configuration example.
**For other configurations**, refer to the [Pre-configured Agent Settings](#pre-configured-agent-settings) table above to see which environment variables are required.
π Click to expand optional API keys
```bash
# API for LLM-as-a-Judge (for benchmark testing, required for benchmark evaluation)
OPENAI_API_KEY=your_openai_key
OPENAI_BASE_URL="https://api.openai.com/v1" # Optional, defaults to OpenAI's API
# API for Open-Source Audio Transcription Tool (for benchmark testing, optional)
WHISPER_MODEL_NAME="openai/whisper-large-v3-turbo"
WHISPER_API_KEY=your_whisper_key
WHISPER_BASE_URL="https://your_whisper_base_url/v1"
# API for Open-Source VQA Tool (for benchmark testing, optional)
VISION_MODEL_NAME="Qwen/Qwen2.5-VL-72B-Instruct"
VISION_API_KEY=your_vision_key
VISION_BASE_URL="https://your_vision_base_url/v1/chat/completions"
# API for Open-Source Reasoning Tool (for benchmark testing, optional)
REASONING_MODEL_NAME="Qwen/Qwen3-235B-A22B-Thinking-2507"
REASONING_API_KEY=your_reasoning_key
REASONING_BASE_URL="https://your_reasoning_base_url/v1/chat/completions"
# API for Claude Sonnet 3.7 as Commercial Tools (optional)
ANTHROPIC_API_KEY=your_anthropic_key
# API for Sogou Search (optional)
TENCENTCLOUD_SECRET_ID=your_tencent_cloud_secret_id
TENCENTCLOUD_SECRET_KEY=your_tencent_cloud_secret_key
# API for Summary LLM (can use small models like Qwen3-14B or GPT-5-Nano)
SUMMARY_LLM_BASE_URL="https://your_summary_llm_base_url/v1/chat/completions"
SUMMARY_LLM_MODEL_NAME=your_summary_llm_model_name # e.g., "Qwen/Qwen3-14B" or "gpt-5-nano"
SUMMARY_LLM_API_KEY=your_summary_llm_api_key
```
### Serve the MiroThinker Agent
#### Option 1 (Recommended): Serve with SGLang or vLLM
Use SGLang to serve MiroThinker agents at port 61002:
```bash
NUM_GPUS=4
PORT=61002
# Downloading agent from HF (v1.5 recommended)
AGENT_PATH=miromind-ai/MiroThinker-v1.5-30B
# Or use v1.0
# AGENT_PATH=miromind-ai/MiroThinker-v1.0-30B
python3 -m sglang.launch_server \
--model-path $AGENT_PATH \
--tp $NUM_GPUS \
--dp 1 \
--host 0.0.0.0 \
--port $PORT \
--trust-remote-code
```
> **π Server URL**: This will start a server at `http://0.0.0.0:$PORT`. Use this as your server base URL (e.g., `http://0.0.0.0:61002/v1`).
#### Option 2: Quantized Light-Weight Options
We also provide comprehensive guidance for serving MiroThinker agents using CPU-optimized and GPU-accelerated quantization techniques, along with detailed analysis and guidelines for deployment with llama.cpp, Ollama, SGLang, and other inference frameworks.
> **π Complete Guide**: See [Deployment Documentation](apps/gradio-demo/) for detailed deployment instructions.
### Run Your First Task
After setting up the environment and starting your server, run `main.py` to test with a default question: *"What is the title of today's arxiv paper in computer science?"*
```bash
cd apps/miroflow-agent
# Using MiroThinker agents (requires your own server)
uv run python main.py llm=qwen-3 agent=mirothinker_v1.5_keep5_max200 llm.base_url=http://localhost:61002/v1
# Or using Claude (requires ANTHROPIC_API_KEY in .env)
uv run python main.py llm=claude-3-7 agent=single_agent_keep5
# Or using GPT-5 (requires OPENAI_API_KEY in .env)
uv run python main.py llm=gpt-5 agent=single_agent_keep5
```
**To customize your question**, edit `main.py` line 32:
```python
task_description = "Your custom question here"
```
The agent will search the web, execute code if needed, and provide an answer with sources.
> **π More details**: See [apps/miroflow-agent/README.md](apps/miroflow-agent/README.md) for available configurations and troubleshooting.
## π Benchmark Evaluation
> For researchers who want to reproduce our benchmark results or evaluate on standard benchmarks.
### Download Benchmark Data
```bash
cd MiroThinker # Back to project root
wget https://huggingface.co/datasets/miromind-ai/MiroFlow-Benchmarks/resolve/main/data_20251115_password_protected.zip
unzip data_20251115_password_protected.zip
# Password: pf4*
rm data_20251115_password_protected.zip
```
### Run Benchmark Evaluation
> **Note:** For MiroThinker v1.5, use `mirothinker_v1.5_keep5_max200` (with context management), `mirothinker_v1.5_keep5_max400` (with context management), or `mirothinker_v1.5` configurations. For v1.0, use `mirothinker_v1.0_keep5` (with context management) or `mirothinker_v1.0` configurations.
**Available Parameters:**
You can customize the evaluation by setting the following environment variables before running the script:
| Parameter | Default | Description |
|:----------|:--------|:------------|
| `LLM_AGENT` | `"MiroThinker-Agents"` | Agent name identifier |
| `BASE_URL` | `"https://your-api.com/v1"` | Base URL of your server |
| `NUM_RUNS` | Varies by benchmark | Number of evaluation runs (3 for most benchmarks, 8 for GAIA/XBench/FutureX/SEAL-0, 32 for AIME2025) |
| `LLM_PROVIDER` | `"qwen"` | LLM provider (e.g., `qwen`, `openai`, `anthropic`) |
| `AGENT_SET` | `"mirothinker_v1.5_keep5_max200"` | Agent configuration (e.g., `mirothinker_v1.5_keep5_max200`, `mirothinker_v1.5_keep5_max400`, `mirothinker_v1.0_keep5`) |
| `MAX_CONTEXT_LENGTH` | `262144` | Maximum context length (256K) |
| `MAX_CONCURRENT` | `10` | Maximum concurrent tasks |
| `PASS_AT_K` | `1` | Pass@K evaluation metric |
| `TEMPERATURE` | `1.0` | Sampling temperature |
| `API_KEY` | `"xxx"` | API key for the server |
**Example Usage:**
```bash
# Navigate to the miroflow-agent directory first
cd apps/miroflow-agent
# Basic usage with v1.5 (recommended)
NUM_RUNS=8 LLM_AGENT="MiroThinker-v1.5-30B" BASE_URL="https://your-api.com/v1" bash scripts/run_evaluate_multiple_runs_gaia-validation-text-103.sh
# Or with v1.0
# NUM_RUNS=8 LLM_AGENT="MiroThinker-v1.0-30B" BASE_URL="https://your-api.com/v1" bash scripts/run_evaluate_multiple_runs_gaia-validation-text-103.sh
# Customize number of runs and agent configuration (v1.5 with context management)
LLM_AGENT="MiroThinker-v1.5-30B" \
BASE_URL="https://your-api.com/v1" \
NUM_RUNS=8 \
AGENT_SET="mirothinker_v1.5_keep5_max200" \
bash scripts/run_evaluate_multiple_runs_gaia-validation-text-103.sh
# Or with v1.0 configuration (with context management)
# LLM_AGENT="MiroThinker-v1.0-30B" \
# BASE_URL="https://your-api.com/v1" \
# NUM_RUNS=8 \
# AGENT_SET="mirothinker_v1.0_keep5" \
# bash scripts/run_evaluate_multiple_runs_gaia-validation-text-103.sh
```
π Click to expand all benchmark commands
> **β οΈ Important for MiroThinker v1.5**: To reproduce our reported results, you must set the correct `AGENT_SET`:
>
> - **BrowseComp & BrowseComp-ZH**: Use `AGENT_SET="mirothinker_v1.5_keep5_max400"`
> - **All other benchmarks**: Use `AGENT_SET="mirothinker_v1.5_keep5_max200"`
```bash
# Navigate to the miroflow-agent directory first
cd apps/miroflow-agent
# HLE
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_hle.sh
# HLE-Text-2158
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_hle-text-2158.sh
# HLE-Text-500
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_hle-text-500.sh
# GAIA-Text-103
NUM_RUNS=8 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_gaia-validation-text-103.sh
# GAIA-Validation (GAIA-Val-165)
NUM_RUNS=8 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_gaia-validation.sh
# BrowseComp-EN (β οΈ use max400)
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max400" bash scripts/run_evaluate_multiple_runs_browsecomp.sh
# BrowseComp-ZH (β οΈ use max400)
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max400" bash scripts/run_evaluate_multiple_runs_browsecomp_zh.sh
# WebWalkerQA
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_webwalkerqa.sh
# XBench-DeepSearch
NUM_RUNS=8 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_xbench_deepsearch.sh
# FRAMES
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_frames.sh
# SEAL-0
NUM_RUNS=8 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_seal-0.sh
# FutureX
NUM_RUNS=8 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_futurex.sh
# AIME2025
NUM_RUNS=32 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_aime2025.sh
# DeepSearchQA
NUM_RUNS=3 LLM_AGENT="xxx" BASE_URL="xxx" AGENT_SET="mirothinker_v1.5_keep5_max200" bash scripts/run_evaluate_multiple_runs_deepsearchqa.sh
```
#### 3. **Monitor evaluation progress**
π Click to expand progress monitoring commands
```bash
# Navigate to the miroflow-agent directory first
cd apps/miroflow-agent
# For HLE
python benchmarks/check_progress/check_progress_hle.py /path/to/evaluation/logs
# For HLE-Text-2158
python benchmarks/check_progress/check_progress_hle-text-2158.py /path/to/evaluation/logs
# For HLE-Text-500
python benchmarks/check_progress/check_progress_hle-text-500.py /path/to/evaluation/logs
# For BrowseComp-EN
python benchmarks/check_progress/check_progress_browsecomp.py /path/to/evaluation/logs
# For BrowseComp-ZH
python benchmarks/check_progress/check_progress_browsecomp_zh.py /path/to/evaluation/logs
# For GAIA-Validation
python benchmarks/check_progress/check_progress_gaia-validation.py /path/to/evaluation/logs
# For GAIA-Text-103
python benchmarks/check_progress/check_progress_gaia-validation-text-103.py /path/to/evaluation/logs
# For WebWalkerQA
python benchmarks/check_progress/check_progress_webwalkerqa.py /path/to/evaluation/logs
# For Frames
python benchmarks/check_progress/check_progress_frames.py /path/to/evaluation/logs
# For XBench-DeepSearch
python benchmarks/check_progress/check_progress_xbench_deepsearch.py /path/to/evaluation/logs
# For SEAL-0
python benchmarks/check_progress/check_progress_seal-0.py /path/to/evaluation/logs
# For AIME2025
python benchmarks/check_progress/check_progress_aime2025.py /path/to/evaluation/logs
# For DeepSearchQA
python benchmarks/check_progress/check_progress_deepsearchqa.py /path/to/evaluation/logs
```
## π¬ Trace Collection
π Click to expand trace collection commands
```bash
cd apps/collect-trace
# Collect Traces for SFT
bash scripts/collect_trace_claude37.sh
bash scripts/collect_trace_gpt5.sh
# Collect Traces for DPO
bash scripts/collect_trace_qwen3.sh
```
## β FAQ & Troubleshooting
### Common Issues
π§ Click to expand troubleshooting guide
#### **Q: Which version should I use?**
**A:** We recommend **MiroThinker v1.5** β with the minimal configuration:
- **v1.5** β: Latest version with 256K context, world-leading performance. Use config (with context management):
- `mirothinker_v1.5_keep5_max200` (up to 200 turns, recommended for most tasks)
- `mirothinker_v1.5_keep5_max400` (up to 400 turns, only used for BrowseComp and BrowseComp-ZH)
#### **Q: How do I get API keys?**
**A:** You need these keys for minimal setup:
- **SERPER_API_KEY**: Get from [Serper.dev](https://serper.dev/) (Google search API)
- **JINA_API_KEY**: Get from [Jina.ai](https://jina.ai/) (Web scraping)
- **E2B_API_KEY**: Get from [E2B.dev](https://e2b.dev/) (Code execution sandbox)
- **SUMMARY_LLM_API_KEY**: Your LLM API credentials (for content summarization). Can be a small model like Qwen3-14B or GPT-5-Nanoβthe choice has minimal impact on performance.
- **OPENAI_API_KEY**: Get from [OpenAI](https://platform.openai.com/) (Required for benchmark evaluation, used for LLM-as-a-Judge)
- **OPENAI_BASE_URL**: Optional, defaults to `https://api.openai.com/v1`. Can be changed to use OpenAI-compatible APIs.
#### **Q: Agent server connection errors**
**A:** Common issues:
- **Check base URL format**: Should end with `/v1` (e.g., `https://your-api.com/v1`)
- **Verify API key**: Ensure `API_KEY` is set correctly in environment or script
- **Check server status**: Make sure your server is running and accessible
- **Network issues**: Verify firewall/network settings allow connections
#### **Q: Evaluation script fails to run**
**A:** Troubleshooting steps:
1. **Check working directory**: Make sure you're in `apps/miroflow-agent` directory
1. **Verify environment**: Run `uv sync` to ensure dependencies are installed
1. **Check .env file**: Ensure all required environment variables are set
1. **Review logs**: Check `logs/` directory for detailed error messages
1. **Verify data path**: Ensure benchmark data is downloaded and in correct location
#### **Q: Out of memory errors**
**A:** Solutions:
- **Reduce context length**: Set `MAX_CONTEXT_LENGTH` to a smaller value (e.g., 131072 for 128K)
- **Use context management with fewer turns**:
- For v1.5: Use `mirothinker_v1.5_keep5_max200` or `mirothinker_v1.5_keep5_max400` (with context management)
- For v1.0: Use `mirothinker_v1.0_keep5` (with context management)
- **Reduce concurrent tasks**: Set `MAX_CONCURRENT` to a smaller number (e.g., 5)
- **Use smaller agents**:
- For v1.5: Try 30B instead of 235B
- For v1.0: Try 8B or 30B instead of 72B
#### **Q: Tool execution errors**
**A:** Common fixes:
- **E2B errors**: Verify `E2B_API_KEY` is valid and account has credits
- **Serper errors**: Check `SERPER_API_KEY` and rate limits
- **Jina errors**: Verify `JINA_API_KEY` and `JINA_BASE_URL` are correct
- **LLM summarization errors**: Check `SUMMARY_LLM_*` variables and agent availability
#### **Q: How to monitor long-running evaluations?**
**A:** Use the progress monitoring scripts:
```bash
cd apps/miroflow-agent
python benchmarks/check_progress/check_progress_.py /path/to/logs
```
The scripts show completion status, elapsed time, and estimated remaining time.
### Getting Help
- π **Documentation**: Check [MiroFlow Tools README](libs/miroflow-tools/README.md) for tool details
- π¬ **Discord**: Join our [Discord community](https://discord.com/invite/GPqEnkzQZd)
- π **Issues**: Report bugs on [GitHub Issues](https://github.com/MiroMindAI/MiroThinker/issues)
- π§ **Contact**: Visit [our website](https://miromind.ai/) for more information
## π License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
## π Acknowledgments
We extend our sincere gratitude to:
- π **Benchmark Contributors** for the comprehensive evaluation datasets
- π **Open Source Community** for the tools and libraries that make this possible
- π₯ **All Contributors** who have helped make MiroThinker better
Join our community and help us build the future of AI agents!
### References
If you find this project useful in your research, please consider citing:
```
@article{miromind2025mirothinker,
title={MiroThinker: Pushing the Performance Boundaries of Open-Source Research Agents via Model, Context, and Interactive Scaling},
author={MiroMind Team and Bai, Song and Bing, Lidong and Chen, Carson and Chen, Guanzheng and Chen, Yuntao and Chen, Zhe and Chen, Ziyi and Dong, Xuan and others},
journal={arXiv preprint arXiv:2511.11793},
year={2025}
}
```
[](https://star-history.com/#MiroMindAI/MiroThinker&Date) 
