{"slug": "show-hn-aibuildai-agent-2-0-an-ai-agent-that-automatically-builds-ai-models", "title": "Show HN: AIBuildAI Agent 2.0 – an AI agent that automatically builds AI models", "summary": "AIBuildAI Agent 2.5 achieved #1 on OpenAI's MLE-Bench with a score of 70.7%, substantially outperforming the agents ranked 2nd through 5th. The AI agent, which automatically builds AI models, also ranked in the top 5.7% of 3,219 teams in Kaggle's TGS Salt Identification Challenge and in the top 6.6% of 4,370 teams in a heart disease prediction competition. Version 2.5 is available to Max users with features including a live build dashboard, run replay, cross-run memory, MCP research tools, and DeepSeek support.", "body_md": "🏆 #1 on OpenAI [MLE-Bench](https://github.com/openai/mle-bench)\n\n[MLE-Bench](https://github.com/openai/mle-bench)\n\n**[6/17/2026]** The**AIBuildAI Agent 2.5** version is made available to[Max users](https://www.aibuildai.io/#products): a live build dashboard, run replay, cross-run memory, MCP research tools, and DeepSeek support.[Release notes and download](https://github.com/aibuildai/AI-Build-AI/releases/tag/v2.5.1).**[5/26/2026]** The**AIBuildAI Agent 2.0** version is made available to[Pro users](https://www.aibuildai.io/#products).**[5/1/2026]** In the[TGS Salt Identification Challenge](https://www.kaggle.com/competitions/tgs-salt-identification-challenge)hosted by Kaggle, the model automatically developed by our AIBuildAI Agent ranked in the top 5.7%. Among 3,219 teams composed of human experts, this performance reaches the level of top-tier human AI experts. Model and code developed by the Agent:[tasks/tgs-salt-identification-challenge](https://github.com/aibuildai/AI-Build-AI/tree/main/tasks/tgs-salt-identification-challenge).**[4/27/2026]** Excited to announce**AIBuildAI Agent 2.0**! It has once again achieved #1 on OpenAI's MLE-Bench, reaching a score of 70.7% and substantially outperforming the agents ranked 2nd through 5th. Compared to version 1.0, Agent 2.0 introduces several technical advancements, which we will detail in an upcoming technical report. The 2.0 version will be available to Pro users soon.**[4/24/2026]** In a[heart disease prediction competition](https://www.kaggle.com/competitions/playground-series-s6e2/overview)hosted by Kaggle, the model automatically developed by our AIBuildAI Agent ranked in the top 6.6%. Among 4,370 teams composed of human experts, this performance reaches the level of top-tier human AI experts. Model and code developed by the agent:[tasks/playground-series-s6e2](https://github.com/aibuildai/AI-Build-AI/tree/main/tasks/playground-series-s6e2).\n\n## demo.mp4\n\nAIBuildAI is an AI agent that automatically builds AI models. Given a task, it runs an agent loop that analyzes the problem, designs models, writes code to implement them, trains them, tunes hyperparameters, evaluates model performance, and iteratively improves the models. By automating the model development workflow, AIBuildAI reduces much of the manual effort required to build AI models.\n\nOn OpenAI [MLE-Bench](https://github.com/openai/mle-bench), AIBuildAI ranked #1, demonstrating strong performance on real-world AI model building tasks.\n\nAIBuildAI requires a **Linux x86_64** machine (Ubuntu 20.04 or newer).\n\nThere are three versions. **V2.5 (Max)** is the current, most capable version. **V2 (Pro)** and **V1 (free)** remain available.\n\n| Version | Plan | To run it |\n|---|---|---|\nV2.5 (Max) — current |\nMax subscription | `aibuildai login` (Max plan) + a model API key (Anthropic or DeepSeek) |\nV2 (Pro) |\nPro subscription | `aibuildai login` (Pro plan) + an Anthropic API key |\nV1 |\nfree | an Anthropic API key (no account) |\n\nSubscriptions are managed at [accounts.aibuildai.io](https://accounts.aibuildai.io); see the account page for the available plans.\n\n-\n**Subscribe.** Create an account at[accounts.aibuildai.io/sign-up](https://accounts.aibuildai.io/sign-up)and switch to the**Max** plan. -\n**Install.**\n\n```\ncurl -fsSL https://github.com/aibuildai/AI-Build-AI/releases/download/v2.5-latest/aibuildai-linux-x86_64.tar.gz | tar xz && ./aibuildai-linux-x86_64-*/install.sh\n```\n\n-\n**Log in**(required before running):\n\n```\naibuildai login      # opens a browser to sign in\naibuildai whoami     # should show an active Max plan\n```\n\n-\n**Set your model-provider API key.** This is separate from the aibuildai subscription above: the subscription (`aibuildai login`\n\n) lets you run the product; this key pays for the AI model calls. aibuildai uses Anthropic (Claude) by default, so set your Anthropic API key:\n\n```\nexport AIBUILDAI_API_KEY=your-anthropic-api-key\n```\n\nDeepSeek is also supported — set a\n\n`deepseek-*`\n\nid in the config's`llm.model`\n\nand put your DeepSeek key in the same`AIBUILDAI_API_KEY`\n\n. -\n**Run.** V2.5 is driven by a YAML config:\n\n```\naibuildai config > task.yaml    # writes a starter config with every field\n# edit task.yaml: set run.task_name, run.data_root, run.instruction, run.playground_root\naibuildai run task.yaml\n```\n\nOther commands:\n\n`aibuildai memorize`\n\n(summarize past runs into memory),`aibuildai replay <run-dir>`\n\n(replay a finished run),`aibuildai --help`\n\n.\n\n-\n**Subscribe** to the**Pro** plan at[accounts.aibuildai.io/sign-up](https://accounts.aibuildai.io/sign-up). -\n**Install.**\n\n```\ncurl -fsSL https://github.com/aibuildai/AI-Build-AI/releases/download/v2.0-latest/aibuildai-linux-x86_64.tar.gz | tar xz && ./aibuildai-linux-x86_64-*/install.sh\n```\n\n-\n**Log in.**\n\n```\naibuildai login\naibuildai whoami     # should show an active Pro plan\n```\n\n-\n**Set credentials.**\n\n```\nexport ANTHROPIC_API_KEY=your-api-key\n```\n\n-\n**Run.** V2 is driven by command-line flags:\n\n```\naibuildai run --task-name <name> --data-dir <path> \\\n  --playground-dir <path> --instruction \"$(cat task.md)\" --no-form\n```\n\nOr run\n\n`aibuildai`\n\nwith no flags to fill in the parameters in an interactive form.\n\nNo account or subscription required.\n\n-\n**Install.**\n\n```\ncurl -fsSL https://github.com/aibuildai/AI-Build-AI/releases/download/v1.0-latest/aibuildai-linux-x86_64.tar.gz | tar xz && ./aibuildai-linux-x86_64-*/install.sh\n```\n\n-\n**Set your Anthropic API key.**\n\n```\nexport ANTHROPIC_API_KEY=your-api-key\n```\n\n-\n**Run** with command-line flags:\n\n```\naibuildai --task-name <name> --data-dir <path> \\\n  --playground-dir <path> --instruction \"$(cat task.md)\" --no-form\n```\n\nOr run\n\n`aibuildai`\n\nwith no flags to fill in the parameters in an interactive form.\n\n**Important:** run the command directly in your terminal. Do not wrap it in a `.sh`\n\n/`.bash`\n\nscript — running it through a script may cause the TUI (Text User Interface) to crash.\n\nAfter a run completes, the output directory usually looks like (structure may slightly vary by task):\n\n```\n├── candidate_1/  candidate_2/  candidate_3/  # Auto-generated training scripts and model checkpoints\n├── checkpoint.pth       # Best model checkpoint\n├── inference.py         # Standalone inference script for the final model\n├── submission.csv       # Test predictions (if test inputs are provided)\n└── progress.pdf         # Visual progress report\n```\n\nThe main outputs of an AIBuildAI run are the model checkpoints and the script `inference.py`\n\n, which runs predictions with the final model on any data. When the task data folder includes unlabeled test inputs, AIBuildAI also writes a predicted-label file `submission.csv`\n\n.\n\nWe provide ready-to-run task markdowns and datasets in the `tasks/`\n\nfolder of this repository. You can also write your own task description and point the run at your own dataset.\n\n```\ngit clone https://github.com/aibuildai/AI-Build-AI.git\n```\n\nThis project is licensed under the MIT License - see the [LICENSE](/aibuildai/AI-Build-AI/blob/main/LICENSE) file for details.\n\n```\n@article{zhang2026aibuildai,\n    title={AIBuildAI: An AI Agent for Automatically Building AI Models},\n    author={Ruiyi Zhang and Peijia Qin and Qi Cao and Li Zhang and Pengtao Xie},\n    year={2026},\n    journal={arXiv},\n    url={https://arxiv.org/abs/2604.14455}\n}\n@article{zhang2026aibuildai2,\n    title={AIBuildAI-2: A Knowledge-Enhanced Agent for Automatically Building AI Models},\n    author={Ruiyi Zhang and Peijia Qin and Qi Cao and Li Zhang and Pengtao Xie},\n    year={2026},\n    journal={arXiv},\n    url={https://arxiv.org/abs/2605.27873}\n}\n```\n\n", "url": "https://wpnews.pro/news/show-hn-aibuildai-agent-2-0-an-ai-agent-that-automatically-builds-ai-models", "canonical_source": "https://github.com/aibuildai/AI-Build-AI", "published_at": "2026-07-21 01:21:24+00:00", "updated_at": "2026-07-21 01:22:52.184951+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "ai-agents", "ai-products", "ai-research"], "entities": ["AIBuildAI", "OpenAI", "MLE-Bench", "Kaggle", "TGS Salt Identification Challenge", "DeepSeek", "Anthropic", "Claude"], "alternates": {"html": "https://wpnews.pro/news/show-hn-aibuildai-agent-2-0-an-ai-agent-that-automatically-builds-ai-models", "markdown": "https://wpnews.pro/news/show-hn-aibuildai-agent-2-0-an-ai-agent-that-automatically-builds-ai-models.md", "text": "https://wpnews.pro/news/show-hn-aibuildai-agent-2-0-an-ai-agent-that-automatically-builds-ai-models.txt", "jsonld": "https://wpnews.pro/news/show-hn-aibuildai-agent-2-0-an-ai-agent-that-automatically-builds-ai-models.jsonld"}}