SEA Workshop (@SEAWorkshop) header
SEA Workshop (@SEAWorkshop) profile picture

Scaling Environments for Agents (SEA) Workshop (NeurIPS 2025)

Joined December 2025
🚀 SEA Workshop is going LIVE TOMORROW! Join us at NeurIPS 2025 for a full day diving into the Scaling Environments for Agents featuring an incredible lineup of speakers and panelists: @egrefen @Mike_A_Merrill @mialon_gregoire @deepaknathani11 @jl_marino @syz0x1 @qhwang3 Anthony G. Cohn, Eric Sommerlade, @fredsala 📍 Upper Level Room 23ABC 🕘 08:00–17:00 Huge thanks to our sponsors: @TheInclusionAI (@AntLingAGI) @SnorkelAI @SonicjobsApp and @VmaxAI 🙌 🔥 Get ready for a day of insights and inspiring conversations!
17
3
37
13,028
SEA Workshop retweeted
SETA got accepted to NeurIPS 2026! 🎉 We released the full training scripts for DeepSeek-V4-Flash, GLM-5.2 (LoRA), Inkling-Small, and Qwen3.8-27B, with both GRPO and PPO, alongside 4,500+ open-source, verifiable terminal RL environments Huge thanks to the @radixark Miles team for helping us scale the experiments! github.com/camel-ai/seta/tre…
🎉 SETA: Scaling Environments for Terminal Agents has been accepted to the #NeurIPS2026 Evaluations & Datasets Track! We release SETA-Env, the largest open-source verifiable terminal RL dataset (4,500+ envs), with full synthesis & training pipelines 🧵 arxiv.org/abs/2607.10891
17
26
5
198
20,830
SEA Workshop retweeted
🚨 CAMEL AI Live Talk this Friday: Dr. Kernel: Reinforcement Learning Done Right for Triton Kernel Generations with Wei Liu Join the CAMEL-AI 🐫 Community to explore Dr. Kernel, a reinforcement learning framework for generating high-performance Triton kernels, enabling scalable GPU-based training and profiling-driven rewards to produce optimized code beyond standard LLM baselines. ⏲️ Time: 27th Feb | GMT 13:00 | EST 8:00 | CST 21:00| 📍 Location: Google Meet 🔗 Register: forms.gle/TAvK5kDcLHU2hdm6A
1
5
1
24
2,850
We’re incredibly excited to launch Open Benchmarks Grants, a new program committing $3M in grants to fund new open source benchmarks advancing agentic AI. We’re partnering up with @HuggingFace, @togethercompute, @PrimeIntellect, Factory HQ, @harborframework, and @PyTorch to support academic and open-source teams as they define and advance the AI frontier.
6
8
71
664,990
Introducing Lab: A full-stack platform for training your own agentic models Build, evaluate and train on your own environments at scale without managing the underlying infrastructure. Giving everyone their own frontier AI lab.
140
293
138
2,865
3,156,621
SEA Workshop retweeted
Open AI and Anthropic just release GPT-5.3-Codex and Opus 4.6 model, terminal capability is now on top of their list evaluating modal capability. But terminal training hits a wall fast: there aren’t enough high-quality environments. In SETA, we just shipped 1,376 validated terminal environments across: SE • sysadmin • security • debugging • networking • DevOps Compatible with Terminal Bench & Harbor. @Mike_A_Merrill @alexgshaw And we’re scaling fast 👀 Find it in: github.com/camel-ai/seta-env or search for seta-env in on harbor registry
4
6
1
41
7,183
SEA Workshop retweeted
🚨 CAMEL-AI Live Talk: We have invited @KeZixuan (Research Scientist at Salesforce AI Research) to share MAS-ZERO: Designing Multi-Agent Systems with Zero Supervision This paper has been accepted to the @SEAWorkshop at NeurIPS 2025, proposing a zero-supervision framework for automatically designing and refining multi-agent systems at inference time, enabling adaptive agent compositions without manual system design. ⏲️ 23 Jan 2026 12:00 GMT / 7:00 EST / 20:00 CST 📍Google meet 🔗Register: forms.gle/wbrmAmtG5zXHbufm8
3
2
16
1,214
🚀 Last December we presented ATLAS at NeurIPS @SEAWorkshop. Now thrilled to share at #AAAI2026 @RealAAAI! ⚠️ RL policy generalization across tasks & levels is hard, especially when some tasks aren't realizable. 🗺️ ATLAS tackles this via autocurricula over tasks AND levels 🧵
1
2
2
18
1,962
SEA Workshop retweeted
It was really fun to meet new people and discuss agent environments. Thanks to the workshop organizers for putting together such a great event! Here is the slide deck from the talk: shuyanzhou.com/assets/slides…
Invited Talk 6 "Towards Future-proof Benchmarks for Digital Agents" from Shuyan Zhou @syz0x1 (Duke University)
3
9
61
7,276
Congratulations also to @_danielfb for winning the best tweet award! Thanks for featuring @SEAWorkshop and for the nice content!
🚀 Thrilled to be presenting ATLAS at the @SEAWorkshop at #NeurIPS2025 this Sunday! Training RL agents to generalize across diverse levels and tasks is hard, especially when some tasks are not always realizable. ATLAS tackles this by creating curricula over tasks AND levels 🧵
1
2
9
928
🏆 Excited to share GEM received the Outstanding Paper Award @SEAWorkshop of #NeurIPS2025 . What a great way to wrap up this amazing neurips journey! Huge thanks to the workshop committee and organizers for the recognition. Grateful for our incredible collaborators and advisors who made this project possible. Thanks to everyone involved! 🎉
Congrats to the following paper authors attaining Outstanding Paper Awards at @SEAWorkshop! GEM: A Gym for Agentic LLMs Zichen Liu, Anya Sims, Keyu Duan, Changyu Chen, Haotian Xu, Simon Yu, Chenmien Tan, Shaopan Xiong, Weixun Wang, Bo Liu, Hao Zhu, Weiyan Shi, Diyi Yang, Wee Sun Lee, Min Lin
6
15
66
9,205
Excellent discussion about closing the agent post-training → evaluation gap @SEAWorkshop today: 1. Scaling complexity: To truly test unseen/real-world capabilities, we need custom environments that scale on specific axes, e.g. tools, open-endedness, dynamic/continuous workloads. 2. Experts-in-the-loop: Domain experts define environment specs and taxonomies—telling us what "hard" actually looks like in their field. 3. Fine-grained verifiers: Verification hooks enable richer reward signals from agent traces, and make agent development more like software engineering. 4. Infra costs: Important to consider infrastructure costs of new harnesses and simulators— these can quickly become a bottleneck if not handled intentionally!
2
2
15
904
The best poster awards go to: 1. Go-Browse: Training Web Agents with Structured Exploration Apurva Gandhi, Graham Neubig 2. Scaling Open-Ended Reasoning to Predict the Future Nikhil Chandak, Shashwat Goel, Ameya Prabhu, Moritz Hardt, Jonas Geiping 🎉Congrats!
2
9
1
21
7,839
Congrats to the following paper authors attaining Outstanding Paper Awards at @SEAWorkshop! GEM: A Gym for Agentic LLMs Zichen Liu, Anya Sims, Keyu Duan, Changyu Chen, Haotian Xu, Simon Yu, Chenmien Tan, Shaopan Xiong, Weixun Wang, Bo Liu, Hao Zhu, Weiyan Shi, Diyi Yang, Wee Sun Lee, Min Lin
3
4
1
22
10,620
Congrats to the following paper authors attaining Outstanding Paper Awards at @SEAWorkshop! RPGBENCH: Evaluating Large Language Models as Role-Playing Game Engines Pengfei Yu, Dongming Shen, Silin Meng, Jaewon Lee, Weisu Yin, Andrea Yaoyun Cui, Zhenlin Xu, Yi Zhu, Xingjian Shi, Mu Li, Alex Smola
3
2
11
637
Would be around the scaling environments for agents workshop #NeurIPS2025 today 🌊 Come say hi and chat about GEM 💎, RL and experience scaling 💡 gem: arxiv.org/pdf/2510.01051
In the era of experience, we're training LLM agents with RL — but something's missing... We miss the good old Gym! So we built 💎GEM: a suite of environments for training LLM 𝚐𝚎𝚗𝚎𝚛𝚊𝚕𝚒𝚜𝚝𝚜. Let’s build the Gym for LLMs, together: axon-rl.notion.site/gem
1
16
1
98
12,683
SEA Workshop retweeted
Now presenting our second poster, “Enabling multi-agent collaboration in knowledge graph environments,” at the @NeurIPSConf Scaling Environments for Agents workshop, wrapping up soon! ft. Yusuf who stopped by to say hi, always nice to catch up with friends at NeurIPS 😁
Presenting our @marinkazitnik lab poster with @lvvittor on “A scalable platform to build the data layer of knowledge graph AI” for the next 1 hour, at the @NeurIPSConf New Perspectives in Graph ML workshop. Come say hi!
4
29
5,523
Come see @getdarshan and @VarunGangal present MEMTRACK at the Scaling Environments for Agents workshop right now at Upper Level 23ABC at #NeurIPS2025 San Diego🔥 We built a RL environment that measures long-term memory and state tracking by putting agents in a workplace with access to real-world productivity tools like slack, linear and git history. Agents must answer questions that reference contexts across platforms and chronological timelines. Best performing model is GPT-5 with 60% correctness score 👀 Full paper: arxiv.org/abs/2510.01353
4
6
33
3,010
SEA Workshop retweeted
Very excited to sponsor the scaling environments workshop
🚀 SEA Workshop is going LIVE TOMORROW! Join us at NeurIPS 2025 for a full day diving into the Scaling Environments for Agents featuring an incredible lineup of speakers and panelists: @egrefen @Mike_A_Merrill @mialon_gregoire @deepaknathani11 @jl_marino @syz0x1 @qhwang3 Anthony G. Cohn, Eric Sommerlade, @fredsala 📍 Upper Level Room 23ABC 🕘 08:00–17:00 Huge thanks to our sponsors: @TheInclusionAI (@AntLingAGI) @SnorkelAI @SonicjobsApp and @VmaxAI 🙌 🔥 Get ready for a day of insights and inspiring conversations!
4
8
1,308
Poster Session 2 is generating massive interest too! Make sure to stop by and discuss the papers with the authors! 📍Upper level room 23 ABC The session concludes at 4:50 PM and will be followed by the closing remark where paper awards will be announced.
1
5
328