mlsanityData to 14 Oct 2026

Seed1.5-Thinking: Advancing Superb Reasoning Models with Reinforcement Learning

arXiv 2504.13914 ↗First posted 10 Apr 2025cs.CLcited by 161

On the map: LLM reasoning → Reasoning benchmarks and evaluation

also in 1 more
Abstract (arXiv)

We introduce Seed1.5-Thinking, capable of reasoning through thinking before responding, resulting in improved performance on a wide range of benchmarks. Seed1.5-Thinking achieves 86.7 on AIME 2024, 55.0 on Codeforces and 77.3 on GPQA, demonstrating excellent reasoning abilities in STEM and coding. Beyond reasoning tasks, the method demonstrates notable generalization across diverse domains. For instance, it surpasses DeepSeek R1 by 8% in win rate on non-reasoning tasks, indicating its broader applicability. Compared to other state-of-the-art reasoning models, Seed1.5-Thinking is a Mixture-of-Experts (MoE) model with a relatively small size, featuring 20B activated and 200B total parameters. As part of our effort to assess generalized reasoning, we develop two internal benchmarks, BeyondAIME and Codeforces, both of which will be publicly released to support future research. Model trial link: https://www.volcengine.com/experience/ark.

Key facts. Cited by 161 papers in the map (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 8 of its 19 references that are in the map. First posted on arXiv 10 Apr 2025.

Builds on

8 of 19 references.

References · 8

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2024HybridFlow: A Flexible and Efficient RLHF Frameworkcited by 2,057
  4. 2016Training Deep Nets with Sublinear Memory Costcited by 848
  5. 2017Ray: A Distributed Framework for Emerging AI Applicationscited by 527
  6. 2025What's Behind PPO's Collapse in Long-CoT? Value Optimization Holds the Secretcited by 98
  7. 2025Exploring Data Scaling Trends and Effects in Reinforcement Learning from Human Feedbackcited by 26
  8. 2025A Unified Pairwise Framework for RLHF: Bridging Generative Reward Modeling and Policy Optimizationcited by 21

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 103; 2026: 58.

161 papers.

Sort:

By area:

  1. 2025MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attentioncited by 263
  2. 2025MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agentcited by 246
  3. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 180
  4. 2025Seed-Prover: Deep and Broad Reasoning for Automated Theorem Provingcited by 101
  5. 2025The Art of Scaling Reinforcement Learning Compute for LLMscited by 101
  6. 2025Soft Adaptive Policy Optimizationcited by 96
  7. 2025Kwai Keye-VL 1.5 Technical Reportcited by 73
  8. 2025WideSearch: Benchmarking Agentic Broad Info-Seekingcited by 69
  9. 2025StreamRL: Scalable, Heterogeneous, and Elastic RL for LLMs with Disaggregated Stream Generationcited by 61
  10. 2025One Token to Fool LLM-as-a-Judgecited by 60
  11. 2025FutureX: An Advanced Live Benchmark for LLM Agents in Future Predictioncited by 59
  12. 2025Reinforcement Learning from Human Feedbackcited by 58
38 more
  1. 2025Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzlescited by 56
  2. 2025rStar2-Agent: Agentic Reasoning Technical Reportcited by 54
  3. 2025Advancing Multimodal Reasoning: From Optimized Cold Start to Staged Reinforcement Learningcited by 53
  4. 2025OneRec-Think: In-Text Reasoning for Generative Recommendationcited by 51
  5. 2025Kwai Keye-VL Technical Reportcited by 47
  6. 2025Harnessing Uncertainty: Entropy-Modulated Policy Gradients for Long-Horizon LLM Agentscited by 46
  7. 2026DeepPlanning: Benchmarking Long-Horizon Agentic Planning with Verifiable Constraintscited by 46
  8. 2025VerIF: Verification Engineering for Reinforcement Learning in Instruction Followingcited by 44
  9. 2025AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Trainingcited by 41
  10. 2025Kimi-Dev: Agentless Training as Skill Prior for SWE-Agentscited by 41
  11. 2025Insights into DeepSeek-V3: Scaling Challenges and Reflections on Hardware for AI Architecturescited by 39
  12. 2025AMO-Bench: Large Language Models Still Struggle in High School Math Competitionscited by 38
  13. 2025Laminar: A Scalable Asynchronous RL Post-Training Frameworkcited by 37
  14. 2025Prolonged Reasoning Is Not All You Need: Certainty-Based Adaptive Routing for Efficient LLM/MLLM Reasoningcited by 35
  15. 2025MathCanvas: Intrinsic Visual Chain-of-Thought for Multimodal Mathematical Reasoningcited by 35
  16. 2025MUA-RL: Multi-turn User-interacting Agent Reinforcement Learning for agentic tool usecited by 32
  17. 2025A Survey on Parallel Text Generation: From Parallel Decoding to Diffusion Language Modelscited by 30
  18. 2025BusterX++: Towards Unified Cross-Modal AI-Generated Content Detection and Explanation with MLLMcited by 29
  19. 2025Scaling up Multi-Turn Off-Policy RL and Multi-Agent Tree Search for LLM Step-Proverscited by 28
  20. 2025Single-stream Policy Optimizationcited by 28
  21. 2025IVEBench: Modern Benchmark Suite for Instruction-Guided Video Editing Assessmentcited by 28
  22. 2025Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agentscited by 24
  23. 2026Think3D: Thinking with Space for Spatial Reasoningcited by 24
  24. 2025From Accuracy to Robustness: A Study of Rule- and Model-based Verifiers in Mathematical Reasoningcited by 22
  25. 2025Introducing LongCat-Flash-Thinking: A Technical Reportcited by 22
  26. 2025Knapsack RL: Unlocking Exploration of LLMs via Optimizing Budget Allocationcited by 22
  27. 2025What Makes a Good Reasoning Chain? Uncovering Structural Patterns in Long Chain-of-Thought Reasoningcited by 21
  28. 2025Rethinking Thinking Tokens: LLMs as Improvement Operatorscited by 21
  29. 2025VerifyBench: Benchmarking Reference-based Reward Systems for Large Language Modelscited by 20
  30. 2025Mathematical Proof as a Litmus Test: Revealing Failure Modes of Advanced Large Reasoning Modelscited by 20
  31. 2025InternBootcamp: Boosting LLM Reasoning with Verifiable Task Scalingcited by 20
  32. 2025Improving RL Exploration for LLM Reasoning through Retrospective Replaycited by 19
  33. 2025Plan Then Action:High-Level Planning Guidance Reinforcement Learning for LLM Reasoningcited by 19
  34. 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
  35. 2025Scaling Law for Quantization-Aware Trainingcited by 17
  36. 2025RBench-V: A Primary Assessment for Visual Reasoning Models with Multi-modal Outputscited by 17
  37. 2025Generalizable End-to-End Tool-Use RL with Synthetic CodeGymcited by 17
  38. 2025AgentMath: Empowering Mathematical Reasoning for Large Language Models via Tool-Augmented Agentcited by 17

50 most cited of 161 All 161 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
  2. 2025Think or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Models
  3. 2025Are DeepSeek R1 And Other Reasoning Models More Faithful?
  4. 2025Seed1.5-VL Technical Report
  5. 2025Brief analysis of DeepSeek R1 and its implications for Generative AI
  6. 2025AM-Thinking-v1: Advancing the Frontier of Reasoning at 32B Scale
  7. 2025Exploring the Effect of Reinforcement Learning on Video Understanding: Insights from SEED-Bench-R1
  8. 2025DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning