{"work":{"id":"e6b75ad5-2877-4168-97c8-710407094d20","openalex_id":"https://openalex.org/W4403625924","doi":"10.1016/j.artmed.2024.103001","arxiv_id":"2501.12948","raw_key":null,"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","authors":null,"authors_text":"DeepSeek-AI","year":2025,"venue":"cs.CL","abstract":"General reasoning represents a long-standing and formidable challenge in artificial intelligence. Recent breakthroughs, exemplified by large language models (LLMs) and chain-of-thought prompting, have achieved considerable success on foundational reasoning tasks. However, this success is heavily contingent upon extensive human-annotated demonstrations, and models' capabilities are still insufficient for more complex problems. Here we show that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning (RL), obviating the need for human-labeled reasoning trajectories. The proposed RL framework facilitates the emergent development of advanced reasoning patterns, such as self-reflection, verification, and dynamic strategy adaptation. Consequently, the trained model achieves superior performance on verifiable tasks such as mathematics, coding competitions, and STEM fields, surpassing its counterparts trained via conventional supervised learning on human demonstrations. Moreover, the emergent reasoning patterns exhibited by these large-scale models can be systematically harnessed to guide and enhance the reasoning capabilities of smaller models.","external_url":"https://arxiv.org/abs/2501.12948","cited_by_count":24,"metadata_source":"pith","metadata_fetched_at":"2026-08-05T02:28:24.338817+00:00","pith_arxiv_id":"2501.12948","created_at":"2026-05-08T17:13:38.764111+00:00","updated_at":"2026-08-05T02:28:24.338817+00:00","title_quality_ok":true,"display_title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","render_title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning"},"hub":{"state":{"work_id":"e6b75ad5-2877-4168-97c8-710407094d20","tier":"mega_hub","tier_reason":"1,000+ Pith inbound or 100,000+ external citations","pith_inbound_count":1939,"external_cited_by_count":24,"distinct_field_count":43,"first_pith_cited_at":"2024-05-01T15:59:00+00:00","last_pith_cited_at":"2026-07-09T16:17:10+00:00","author_build_status":"needed","summary_status":"needed","contexts_status":"needed","graph_status":"needed","ask_index_status":"needed","reader_status":"needed","recognition_status":"needed","updated_at":"2026-08-23T07:09:32.911275+00:00","tier_text":"mega_hub"},"tier":"mega_hub","role_counts":[{"context_role":"background","n":272},{"context_role":"method","n":48},{"context_role":"baseline","n":28},{"context_role":"dataset","n":12},{"context_role":"other","n":6}],"polarity_counts":[{"context_polarity":"background","n":259},{"context_polarity":"use_method","n":45},{"context_polarity":"baseline","n":28},{"context_polarity":"unclear","n":18},{"context_polarity":"use_dataset","n":11},{"context_polarity":"support","n":4},{"context_polarity":"extend","n":1}],"runs":{"ask_index":{"job_type":"ask_index","status":"succeeded","result":{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","claims":[{"claim_text":"General reasoning represents a long-standing and formidable challenge in artificial intelligence. Recent breakthroughs, exemplified by large language models (LLMs) and chain-of-thought prompting, have achieved considerable success on foundational reasoning tasks. However, this success is heavily contingent upon extensive human-annotated demonstrations, and models' capabilities are still insufficient for more complex problems. Here we show that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning (RL), obviating the need for human-labeled reasoning trajectorie","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-13T17:53:36.998847+00:00"},"author_expand":{"job_type":"author_expand","status":"succeeded","result":{"authors_linked":[{"id":"b3d3bc38-c7e6-4554-ab1a-a4b32a8299c8","orcid":null,"display_name":"DeepSeek-AI"}]},"error":null,"updated_at":"2026-05-13T17:24:04.935515+00:00"},"context_extract":{"job_type":"context_extract","status":"succeeded","result":{"enqueued_papers":25},"error":null,"updated_at":"2026-05-13T17:53:36.992694+00:00"},"graph_features":{"job_type":"graph_features","status":"succeeded","result":{"co_cited":[{"title":"DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models","work_id":"c5006563-f3ec-438a-9e35-b7b484f34828","shared_citers":241},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":211},{"title":"Proximal Policy Optimization Algorithms","work_id":"240c67fe-d14d-4520-91c1-38a4e272ca19","shared_citers":137},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":114},{"title":"The Llama 3 Herd of Models","work_id":"1549a635-88af-4ac1-acfe-51ae7bb53345","shared_citers":114},{"title":"DAPO: An Open-Source LLM Reinforcement Learning System at Scale","work_id":"64019d00-0b11-4bbd-b173-b46c8fad0157","shared_citers":113},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":104},{"title":"OpenAI o1 System Card","work_id":"68d3c334-0fc9-49e3-b7b0-a69afae933e2","shared_citers":103},{"title":"Qwen2.5 Technical Report","work_id":"d8432992-4980-4a81-85c7-9fa2c2b87f85","shared_citers":77},{"title":"GPT-4o System Card","work_id":"f37bf1c7-4964-4e56-9762-d20da8d9009f","shared_citers":73},{"title":"Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities","work_id":"008df105-2fdd-45d8-857a-8e35868aecb6","shared_citers":72},{"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","shared_citers":72},{"title":"DeepSeek-V3 Technical Report","work_id":"57d2791d-2219-4c31-a077-afc04b12a75c","shared_citers":67},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":61},{"title":"Measuring Mathematical Problem Solving With the MATH Dataset","work_id":"50652ac6-fb7c-4675-a2c2-159c241feb17","shared_citers":56},{"title":"Group Sequence Policy Optimization","work_id":"3a98b53b-9f52-4d95-adf7-89353c0a9a65","shared_citers":51},{"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","shared_citers":51},{"title":"Kimi k1.5: Scaling Reinforcement Learning with LLMs","work_id":"bff96ab1-bd6a-4585-be23-74fdb51969c7","shared_citers":47},{"title":"Understanding R1-Zero-Like Training: A Critical Perspective","work_id":"ec354f3b-9484-4a0c-94c8-92d4d0260835","shared_citers":45},{"title":"Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters","work_id":"a8d50b24-bdf5-46ed-bc4f-2927dfd81f1d","shared_citers":41},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":40},{"title":"LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code","work_id":"ea9e51ce-1e75-4182-92d8-4d25f70d2ee4","shared_citers":35},{"title":"Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution","work_id":"8abcfe4f-e0fb-44b7-9123-448fac95f90a","shared_citers":35},{"title":"Self-Consistency Improves Chain of Thought Reasoning in Language Models","work_id":"8c6d5a6b-b5cc-4105-9c84-9c34bb9375bb","shared_citers":34}],"time_series":[{"n":2,"year":2024},{"n":34,"year":2025},{"n":563,"year":2026}]},"error":null,"updated_at":"2026-05-13T17:25:54.879041+00:00"},"identity_refresh":{"job_type":"identity_refresh","status":"succeeded","result":{"fixed":1,"items":[{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","resolver":"local_arxiv","confidence":0.98,"old_work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e"}],"errors":[],"attempted":1},"error":null,"updated_at":"2026-05-13T17:53:36.131917+00:00"},"reader_index":{"job_type":"reader_index","status":"succeeded","result":{"note":"annotated reader requires full-text/OA fetch; shell is wired for mega hubs","status":"reader queued"},"error":null,"updated_at":"2026-05-19T02:51:43.032474+00:00"},"recognition_alignment":{"job_type":"recognition_alignment","status":"succeeded","result":{"modules":["IndisputableMonolith.Cognition.AnalogicalReasoningFromJCost","IndisputableMonolith.Cognition.AnimalZComplexityBound","IndisputableMonolith.Foundation.VoiceForcing","IndisputableMonolith.Education.PedagogyModelsFromConfigDim","IndisputableMonolith.Cosmology.CosmologicalConstant","IndisputableMonolith.Core.ConstantsAndPatterns","IndisputableMonolith.Materials.RoomTSuperconductorCandidate","IndisputableMonolith.Common.CanonicalJBand"],"query_chars":1270},"error":null,"updated_at":"2026-05-19T02:51:43.029049+00:00"},"role_polarity":{"job_type":"role_polarity","status":"succeeded","result":{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","claims":[{"claim_text":"General reasoning represents a long-standing and formidable challenge in artificial intelligence. Recent breakthroughs, exemplified by large language models (LLMs) and chain-of-thought prompting, have achieved considerable success on foundational reasoning tasks. However, this success is heavily contingent upon extensive human-annotated demonstrations, and models' capabilities are still insufficient for more complex problems. Here we show that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning (RL), obviating the need for human-labeled reasoning trajectorie","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-13T17:53:36.995986+00:00"},"summary_claims":{"job_type":"summary_claims","status":"succeeded","result":{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","claims":[{"claim_text":"General reasoning represents a long-standing and formidable challenge in artificial intelligence. Recent breakthroughs, exemplified by large language models (LLMs) and chain-of-thought prompting, have achieved considerable success on foundational reasoning tasks. However, this success is heavily contingent upon extensive human-annotated demonstrations, and models' capabilities are still insufficient for more complex problems. Here we show that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning (RL), obviating the need for human-labeled reasoning trajectorie","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-13T17:25:52.711836+00:00"}},"summary":{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","claims":[{"claim_text":"General reasoning represents a long-standing and formidable challenge in artificial intelligence. Recent breakthroughs, exemplified by large language models (LLMs) and chain-of-thought prompting, have achieved considerable success on foundational reasoning tasks. However, this success is heavily contingent upon extensive human-annotated demonstrations, and models' capabilities are still insufficient for more complex problems. Here we show that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning (RL), obviating the need for human-labeled reasoning trajectorie","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning because it crossed a citation-hub threshold.","role_counts":[]},"graph":{"co_cited":[{"title":"DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models","work_id":"c5006563-f3ec-438a-9e35-b7b484f34828","shared_citers":241},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":211},{"title":"Proximal Policy Optimization Algorithms","work_id":"240c67fe-d14d-4520-91c1-38a4e272ca19","shared_citers":137},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":114},{"title":"The Llama 3 Herd of Models","work_id":"1549a635-88af-4ac1-acfe-51ae7bb53345","shared_citers":114},{"title":"DAPO: An Open-Source LLM Reinforcement Learning System at Scale","work_id":"64019d00-0b11-4bbd-b173-b46c8fad0157","shared_citers":113},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":104},{"title":"OpenAI o1 System Card","work_id":"68d3c334-0fc9-49e3-b7b0-a69afae933e2","shared_citers":103},{"title":"Qwen2.5 Technical Report","work_id":"d8432992-4980-4a81-85c7-9fa2c2b87f85","shared_citers":77},{"title":"GPT-4o System Card","work_id":"f37bf1c7-4964-4e56-9762-d20da8d9009f","shared_citers":73},{"title":"Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities","work_id":"008df105-2fdd-45d8-857a-8e35868aecb6","shared_citers":72},{"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","shared_citers":72},{"title":"DeepSeek-V3 Technical Report","work_id":"57d2791d-2219-4c31-a077-afc04b12a75c","shared_citers":67},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":61},{"title":"Measuring Mathematical Problem Solving With the MATH Dataset","work_id":"50652ac6-fb7c-4675-a2c2-159c241feb17","shared_citers":56},{"title":"Group Sequence Policy Optimization","work_id":"3a98b53b-9f52-4d95-adf7-89353c0a9a65","shared_citers":51},{"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","shared_citers":51},{"title":"Kimi k1.5: Scaling Reinforcement Learning with LLMs","work_id":"bff96ab1-bd6a-4585-be23-74fdb51969c7","shared_citers":47},{"title":"Understanding R1-Zero-Like Training: A Critical Perspective","work_id":"ec354f3b-9484-4a0c-94c8-92d4d0260835","shared_citers":45},{"title":"Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters","work_id":"a8d50b24-bdf5-46ed-bc4f-2927dfd81f1d","shared_citers":41},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":40},{"title":"LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code","work_id":"ea9e51ce-1e75-4182-92d8-4d25f70d2ee4","shared_citers":35},{"title":"Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution","work_id":"8abcfe4f-e0fb-44b7-9123-448fac95f90a","shared_citers":35},{"title":"Self-Consistency Improves Chain of Thought Reasoning in Language Models","work_id":"8c6d5a6b-b5cc-4105-9c84-9c34bb9375bb","shared_citers":34}],"time_series":[{"n":2,"year":2024},{"n":34,"year":2025},{"n":563,"year":2026}]},"authors":[{"id":"b3d3bc38-c7e6-4554-ab1a-a4b32a8299c8","orcid":null,"display_name":"DeepSeek-AI","source":"manual","import_confidence":0.72}]}}