OpenThoughts-Agent-SFT-31.6K

<p align="center"> <img src="https://huggingface.co/datasets/open-thoughts/OpenThoughts1-Agent-SFT/resolve/main/ota-logo.png" width="50%"> </p> <p align="center"> <a href="https://www.openthoughts.ai/blog/agent" style="margin-right: 24px;">Project</a> | <a href="https://github.com/open-thoughts/OpenThoughts-Agent" style="margin-right: 24px; margin-left: 24px;">Code</a> | <a href="https://huggingface.co/collections/open-thoughts/openthinker-agent" style="margin-left: 24px;">Collection</a> </p> # OpenThoughts-Agent-SFT-31.6K **OpenThoughts-Agent** is an open-source effort to curate the best datasets for training agents. Our release includes [datasets](https://huggingface.co/collections/open-thoughts/openthinker-agent), [models](https://huggingface.co/collections/open-thoughts/openthinker-agent) and our [research codebase](https://github.com/open-thoughts/OpenThoughts-Agent). [OpenThoughts-Agent-SFT-31.6K](https://huggingface.co/datasets/open-thoughts/OpenThoughts-Agent-SFT-31.6K) is the **31,600-example** point of the OpenThoughts-Agent SFT scaling ladder (sizes 316 / 1K / 3.16K / 10K / 31.6K / 100K). It contains (task, agent-trajectory) pairs used to fine-tune [OpenThinkerAgent-8B-SFT-31.6K](https://huggingface.co/open-thoughts/OpenThinkerAgent-8B-SFT-31.6K) and [OpenThinkerAgent-32B-SFT-31.6K](https://huggingface.co/open-thoughts/OpenThinkerAgent-32B-SFT-31.6K). The 100K set is the final OpenThoughts-Agent SFT dataset described in the paper. - **Homepage:** https://www.openthoughts.ai/blog/agent - **Repository:** https://github.com/open-thoughts/OpenThoughts-Agent # Data Tasks are drawn from the **Top-4** task sources identified by our ablations: **SWE-Smith**, **StackExchange-SuperUser**, **StackExchange-Tezos** (synthetically augmented to expand task diversity), and **IssueTasks**. Agentic trajectories are generated by **GLM-4.7-AWQ** acting as the teacher in the **terminus-2** harness inside Daytona sandboxes, then filtered to traces with **at least 5 model turns**. | Field | Description | | --- | --- | | `conversations` | the multi-turn agent trajectory (role/content messages) | | `task` | the task description given to the agent | | `trace_source` | originating task source (swesmith / superuser / tezos / issue-tasks) | | `agent`, `model`, `model_provider` | rollout harness and teacher metadata | | `result`, `episode`, `run_id`, `trial_name`, `date` | rollout bookkeeping | - **Rows:** 31,600 - **Teacher:** GLM-4.7-AWQ &middot; **Harness:** terminus-2 # Links - 🌐 [OpenThoughts-Agent project page](https://www.openthoughts.ai/blog/agent) - 💻 [OpenThoughts-Agent GitHub repository](https://github.com/open-thoughts/OpenThoughts-Agent) - 📚 [OpenThinker-Agent collection](https://huggingface.co/collections/open-thoughts/openthinker-agent) - 🤖 [OpenThinkerAgent-32B-SFT-31.6K model](https://huggingface.co/open-thoughts/OpenThinkerAgent-32B-SFT-31.6K) - 🤖 [OpenThinkerAgent-8B-SFT-31.6K model](https://huggingface.co/open-thoughts/OpenThinkerAgent-8B-SFT-31.6K) # Citation ``` @misc{openthoughts-agent, author = {Team, OpenThoughts-Agent}, title = {{OpenThoughts-Agent: Data Recipes for Agentic Models}}, howpublished = {https://www.openthoughts.ai/blog/agent}, year = {2026} } ```