axion1_0
Frontier-level agent capabilities with nothing more than high-quality data and a simple training recipe: 2 SFT + RL, no mid-training/DPO/OPD. Full recipe released.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Check it deeper on your machine
It has changed since its scan on 11 Nov 2026.
As a Claude Code plugin, install it and then ask Claude Code:
/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/bytedance/axion1_0 with scanskillsafety
Or from a terminal, clone it and run it:
git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/bytedance/axion1_0'
It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.
Badge for your README
Markdown
[](/listings/bytedance--axion1-0)HTML
<a href="/listings/bytedance--axion1-0"><img src="/api/listings/bytedance--axion1-0/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Clone the repository and follow the [training guide](axion/training/README.md) for environment setup, two-stage SFT and checkpoint export. SFT uses Swift with Megatron-FSDP; RL uses verl. They implement successive stages of the same training pipeline.
git clone https://github.com/bytedance/axion1_0.git cd axion1_0
axion/
training/
sft/ Two-stage SFT, runtime patches and checkpoint export
rl/ Agent rollouts, rewards, tools and RL configuration
datagen/ Stage I and Stage II data construction
evaluation/ Benchmark preparation and grading
third_party/
swift/ SFT backend
verl/ RL backend
tests/ Regression and packaging checksThe 122B pipeline consists of two-stage SFT followed by asynchronous GRPO. The training guides cover single-host and multi-host layouts, required services, checkpoint export, and runtime requirements.
For data construction and evaluation, use a separate Python 3.11+ environment:
python -m pip install -e ".[data,evaluation]"
4.1 Inference
Export a native SFT checkpoint to Hugging Face format before serving it. The [inference guide](axion/evaluation/README.md#serve-a-checkpoint) provides a vLLM launch command and an OpenAI-compatible request example. Agent evaluation also requires a tool-execution harness.
4.2 Documentation
| Guide | Contents | |:---|:---| | [Stage I data](axion/datagen/stage1/README.md) | Problem generation, rewriting, answer verification, and trajectory filtering | | [Stage II data](axion/datagen/stage2/README.md) | Database generators, question filtering, scoring, and trajectory processing | | [SFT](axion/training/sft/README.md) | Runtime setup, data requirements, two-stage training, and checkpoint export | | [RL](axion/training/rl/README.md) | Paper recipe, task selection, services, and launch | | [Evaluation](axion/evaluation/README.md) | Dataset preparation and grading for FrontierScience, HiPhO, HLE-130, and SEAL-0 |
Evaluation scripts prepare datasets and grade rollouts collected with an external agent harness. The evaluation guide documents the required format.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 210 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.
- <div align="center">
- <picture>
- <source media="(prefers-color-scheme: dark)" srcset="assets/axion-logo-dark.svg">
- <img src="assets/axion-logo.svg" width="260" alt="Axion 1.0">
- </picture>
- <h1>Open Agent Foundation Models, Simplified</h1>
- <p>
- <a href="https://axion1-0.github.io/"><img src="https://img.shields.io/badge/Project-Website-2563EB?style=flat-square" alt="Project website"></a>
- <a href="https://axion1-0.github.io/paper/axion.pdf"><img src="https://img.shields.io/badge/Read-Technical_Report-0F766E?style=flat-square" alt="Technical report"></a>
- <a href="https://huggingface.co/datasets/Axion1-0/axion-training-data"><img src="https://img.shields.io/badge/Data-Hugging_Face-0F766E?style=flat-square" alt="Training data on Hugging Face"></a>
- <a href="https://huggingface.co/Axion1-0/Axion-1.0-35B"><img src="https://img.shields.io/badge/Model-35B_Reproduction-2563EB?style=flat-square" alt="Axion 1.0-35B reproduction model on Hugging Face"></a>
- <a href="LICENSE"><img src="https://img.shields.io/badge/License-Apache_2.0-2563EB?style=flat-square" alt="Apache 2.0 license"></a>
- </p>
- <p>
- <a href="https://jyqhahah.github.io/">Yuqi Jia</a><sup>1</sup> · <a href="https://sunsmarterjie.github.io/">Yunjie Tian</a><sup>2,†</sup> · <a href="https://www.linkedin.com/in/jiayu-zheng-5aa231250/">Jiayu Zheng</a><sup>2</sup> · <a href="https://joliang17.github.io/">Yijun Liang</a><sup>3</sup> · <a href="https://jingchensun.github.io/">Jingchen Sun</a><sup>4</sup> · <a href="https://www.linkedin.com/in/mengyu-zhao-7ab19274/">Mengyu Zhao</a><sup>2</sup> · Di Fu<sup>2</sup>
- </p>
- <p>
- <sup>1</sup> Duke University <sup>2</sup> ByteDance <sup>3</sup> University of Maryland, College Park <sup>4</sup> University at Buffalo<br>
- <sup>†</sup> Project Lead
- </p>
- </div>
- ## 1. Highlights
- - **Three-stage training.** Two stages of supervised fine-tuning (SFT) followed
- by reinforcement learning (RL), without dedicated mid-training, DPO, or
- on-policy distillation.
- - **Scientific reasoning and tool use.** Axion 1.0-122B scores **59.0** on full
- HLE with restricted website access, **87.0** on FrontierScience Olympiad, and
- **66.7** on FrontierScience Research. Axion 1.0-35B scores **62.2** on SEAL-0.
- - **Verifiable training data.** Reasoning problems are strengthened using
- feedback from the base model. Tasks requiring tools are constructed from
- public data sources with programmatically determined reference answers.
- - **Open training code.** The repository includes data generation and
- filtering, prompts, two-stage SFT, asynchronous RL, and benchmark preparation
- and grading.
- ## 2. Results
- <p align="center">
- <a href="assets/benchmark-results.png">
- <img src="assets/benchmark-results.png" width="100%" alt="Axion 1.0 compared with selected open and proprietary models on HLE, BioMysteryBench, SEAL-0, HiPhO, and both FrontierScience splits">
- </a>
- </p>
- Selected results from Table 1 of the
- [technical report](https://axion1-0.github.io/paper/axion.pdf#page=9):
- | Model | HLE | BMB | SEAL-0 | HiPhO | FS Olympiad | FS Research |
- |:---|---:|---:|---:|---:|---:|---:|
- | Agents-A1 | 47.6 | — | 56.4 | 46.4 | 79.0 | 40.0 |
- | Apodex 1.1 (Agent Team) | 56.1 | 35.3 | — | — | — | 63.3 |
- | GPT-5.6 Sol | — | 47.1* | 53.2* | 42.9* | 79.0* | 63.3* |
- | Gemini 3.1 Pro | 51.4 | 35.3* | 53.2* | 38.5* | 85.0* | 26.7* |
- | Claude Fable 5 | 63.8 | 46.1 | — | — | — | — |
- | **Axion 1.0-35B** | **52.2** | **35.3** | **62.2** | **45.7** | **86.0** | **56.7** |
- | **Axion 1.0-122B** | **59.0** | **41.2** | **58.6** | **46.2** | **87.0** | **66.7** |
- BMB denotes the human-difficult subset of BioMysteryBench; FS denotes
- FrontierScience. * marks baselines evaluated by the Axion team; other baseline
- scores are from official reports. — denotes an unreported result.
- ### 2.1 Models
- | Model | Base model | Total / active parameters |
- |:---|:---|:---:|
- | Axion 1.0-35B | Qwen3.6-35B-A3B | 35B / 3B |
- | Axion 1.0-122B | Qwen3.5-122B-A10B | 122B / 10B |
- Both models in the paper use SFT and RL. The paper also reports 4B and 9B
- variants trained with SFT only.
- We encourage users to try the
- [Axion 1.0-35B community reproduction](https://huggingface.co/Axion1-0/Axion-1.0-35B),
- announced in [issue #1](https://github.com/bytedance/axion1_0/issues/1).
- This is a community release, not an official Axion checkpoint.
- See the model card for reproduction results and inference instructions.
- Training data is available on
- [Hugging Face](https://huggingface.co/datasets/Axion1-0/axion-training-data).
- ## 3. Training Recipe
- <p align="center">
- <a href="assets/data-construction.svg">
- <img src="assets/data-construction.svg" width="100%" alt="Two-stage data construction: model-aware strengthening of self-contained reasoning problems, followed by deterministic construction and verification of tool-dependent tasks">
- </a>
- </p>
- ### 3.1 Stage I: Self-Contained Reasoning
- Stage I uses **10,819 trajectories** of scientific and mathematical reasoning.
- Public seed problems are rewritten into harder questions. Problems solved by
- the base model are further strengthened through composition, parameterization,
- range shifts, or case analysis, with reference answers checked after each
- rewrite. Solvers may use Python, but not search or webpage retrieval.
- ### 3.2 Stage II: Tool-Dependent Reasoning
- Stage II uses **7,544 trajectories**, approximately **312M tokens**. Programs
- filter records, select a unique entity, and compute a reference answer before
- an LLM formulates the question. Solvers must find the required evidence through
- tools; construction data and answers are withheld during solving. Questions
- solvable without retrieval are removed.
- Generators cover individual entities, ranked lists, and chains across
- databases, including the NASA Exoplanet Archive, OurAirports, OpenAlex,
- UniProt, and Ensembl.
- ### 3.3 Reinforcement Learning
- Starting from the Stage II checkpoint, **asynchronous GRPO** trains on agent
- interactions with search, webpage visits, scholarly retrieval, and Python.
- Correct final answers receive full reward; verified intermediate progress can
- receive partial credit. Context condensation divides long rollouts into
- sessions for training.
- ### 3.4 Training Infrastructure
- The paper's SFT runs use a 128K context window with FSDP, expert parallelism,
- and context parallelism. RL separates rollout workers, trainers, and auxiliary
- services into distinct GPU pools. Token-In-Token-Out (TITO) preserves token
- IDs, and Rollout Routing Replay (R3) aligns MoE expert choices between rollout
- and training.
- Two-stage SFT costs reported in Section 5.1:
- | Model | GPUs | Stage I | Stage II | Total GPU-hours |
- |:---|---:|---:|---:|---:|
- | Axion 1.0-35B | 32 | 1h 02m | 1h 46m | ≈90 |
- | Axion 1.0-122B | 32 | 3h 32m | 3h 58m | 240 |
- These costs are reported in the paper. The released 122B SFT recipe uses
- 32 GPUs, a 128K context window, and a global batch of 8. Its saved training
- arguments record micro-batch size 1; Section 5.1 lists 2. The released
- [SFT recipe](axion/training/sft/README.md) preserves the saved training arguments.
- ## 4. Getting Started
- Clone the repository and follow the [training guide](axion/training/README.md)
- for environment setup, two-stage SFT and checkpoint export. SFT uses Swift with
- Megatron-FSDP; RL uses verl. They implement successive stages of the same
- training pipeline.
- ```bash
- git clone https://github.com/bytedance/axion1_0.git
- cd axion1_0
- ```
- ```text
- axion/
- training/
- sft/ Two-stage SFT, runtime patches and checkpoint export
- rl/ Agent rollouts, rewards, tools and RL configuration
- datagen/ Stage I and Stage II data construction
- evaluation/ Benchmark preparation and grading
- third_party/
- swift/ SFT backend
- verl/ RL backend
- tests/ Regression and packaging checks
- ```
- The 122B pipeline consists of two-stage SFT followed by asynchronous GRPO.
- The training guides cover single-host and multi-host layouts, required services,
- checkpoint export, and runtime requirements.
- For data construction and evaluation, use a separate Python 3.11+ environment:
- ```bash
- python -m pip install -e ".[data,evaluation]"
- ```
- ### 4.1 Inference
- Export a native SFT checkpoint to Hugging Face format before serving it. The
- [inference guide](axion/evaluation/README.md#serve-a-checkpoint) provides a vLLM
- launch command and an OpenAI-compatible request example. Agent evaluation
- also requires a tool-execution harness.
- ### 4.2 Documentation
- | Guide | Contents |
- |:---|:---|
- | [Stage I data](axion/datagen/stage1/README.md) | Problem generation, rewriting, answer verification, and trajectory filtering |
- | [Stage II data](axion/datagen/stage2/README.md) | Database generators, question filtering, scoring, and trajectory processing |
- | [SFT](axion/training/sft/README.md) | Runtime setup, data requirements, two-stage training, and checkpoint export |
- | [RL](axion/training/rl/README.md) | Paper recipe, task selection, services, and launch |
- | [Evaluation](axion/evaluation/README.md) | Dataset preparation and grading for FrontierScience, HiPhO, HLE-130, and SEAL-0 |
- Evaluation scripts prepare datasets and grade rollouts collected with an
- external agent harness. The evaluation guide documents the required format.
- ## 5. License
- [Apache 2.0](LICENSE).
- ## 6. Citation
- ```bibtex
- @article{jia2026axion,
- title={Axion 1.0: Open Agent Foundation Models, Simplified},
- author={Jia, Yuqi and Tian, Yunjie and Zheng, Jiayu and Liang, Yijun and Sun, Jingchen and Zhao, Mengyu and Fu, Di},
- journal={arXiv preprint arXiv:2611.03233},
- year={2026}
- }
- ```
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.