# Together AI — DSGym: A holistic framework for evaluating and training data science agents

- Company: Together AI (together.ai)
- Announced: 2026-01-26T00:00:00+00:00
- Category: research-paper
- Subject: Inference platform
- Models affected: Qwen3-4B-DSGym-SFT-2k, GPT-5.1, GPT-5, GPT-4o, Claude Sonnet 4.5, Claude Sonnet 4, Qwen3 235B Instruct, Qwen3-Coder 480B, Kimi K2 Instruct, GPT-OSS-120B, Deepseek-v3.1, Qwen2.5-7B-Instruct, Qwen3-4B-Instruct
- Source: https://www.together.ai/blog/dsgym
- Record: https://forck.live/items/2386-dsgym-a-holistic-framework-for-evaluating-and-training-data-science-agents

Together AI introduces DSGym, a unified framework for evaluating and training data science agents. It integrates existing benchmarks and adds novel scientific analysis tasks (DSBio, 90 bioinformatics tasks) and Kaggle competitions (DSPredict, 92 competitions). Using DSGym, the team trained a 4B model (Qwen3-4B-DSGym-SFT-2k) on 2,000 synthetic query-trajectory pairs, achieving state-of-the-art performance among open-source models on general data science benchmarks.

## Evidence

Verbatim from https://www.together.ai/blog/dsgym:

> We address these limitations by introducing DSGym, an integrated framework for evaluating and training data science agents in self-contained execution environments. Using DSGym, we trained a state-of-the-art open-source data science agent.

---

Record: https://forck.live/items/2386-dsgym-a-holistic-framework-for-evaluating-and-training-data-science-agents
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md