🐦 Twitter Post Details

Viewing enriched Twitter post

@dair_ai

// Multi-Agent Self-Evolution for LLM Reasoning // Most self-play methods for LLM reasoning lack explicit planning and quality control. This leads to unstable training on complex multi-step tasks. New research introduces a cleaner closed-loop approach. SAGE co-evolves four specialized agents from a single LLM backbone using only 500 seed examples: a Challenger generates increasingly harder tasks, a Planner structures step-by-step strategies, a Solver produces answers verified externally, and a Critic scores and filters both questions and plans to prevent curriculum drift. Why does it matter? SAGE achieves consistent gains across model scales with minimal data. That's very desirable. On Qwen-2.5-7B, it improves OOD performance by +4.2% while maintaining in-distribution accuracy, outperforming both Absolute Zero Reasoning and Multi-Agent Evolve baselines across code and math benchmarks. Paper: https://t.co/8Zn41OBIra Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

View on Twitter

📊 Media Metadata

{
  "media": [
    {
      "type": "photo",
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2037548967366738131/media_0.jpg",
      "filename": "media_0.jpg"
    },
    {
      "type": "photo",
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2037548967366738131/media_1.png",
      "filename": "media_1.png"
    }
  ],
  "processed_at": "2026-03-27T15:16:56.831255",
  "pipeline_version": "2.0"
}

🔧 Raw API Response

{
  "type": "tweet",
  "id": "2037548967366738131",
  "url": "https://x.com/dair_ai/status/2037548967366738131",
  "twitterUrl": "https://twitter.com/dair_ai/status/2037548967366738131",
  "text": "// Multi-Agent Self-Evolution for LLM Reasoning //\n\nMost self-play methods for LLM reasoning lack explicit planning and quality control.\n\nThis leads to unstable training on complex multi-step tasks.\n\nNew research introduces a cleaner closed-loop approach.\n\nSAGE co-evolves four specialized agents from a single LLM backbone using only 500 seed examples: a Challenger generates increasingly harder tasks, a Planner structures step-by-step strategies, a Solver produces answers verified externally, and a Critic scores and filters both questions and plans to prevent curriculum drift.\n\nWhy does it matter?\n\nSAGE achieves consistent gains across model scales with minimal data. That's very desirable.\n\nOn Qwen-2.5-7B, it improves OOD performance by +4.2% while maintaining in-distribution accuracy, outperforming both Absolute Zero Reasoning and Multi-Agent Evolve baselines across code and math benchmarks.\n\nPaper: https://t.co/8Zn41OBIra\n\nLearn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c",
  "source": "Twitter for iPhone",
  "retweetCount": 0,
  "replyCount": 0,
  "likeCount": 1,
  "quoteCount": 0,
  "viewCount": 46,
  "createdAt": "Fri Mar 27 15:15:04 +0000 2026",
  "lang": "en",
  "bookmarkCount": 1,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2037548967366738131",
  "displayTextRange": [
    0,
    277
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "dair_ai",
    "url": "https://x.com/dair_ai",
    "twitterUrl": "https://twitter.com/dair_ai",
    "id": "889050642903293953",
    "name": "DAIR.AI",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1643277398522187778/31dedbLo_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/889050642903293953/1773242460",
    "description": "",
    "location": "",
    "followers": 92560,
    "following": 1,
    "status": "",
    "canDm": true,
    "canMediaTag": true,
    "createdAt": "Sun Jul 23 09:12:45 +0000 2017",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 4282,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 180,
    "statusesCount": 3039,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [
      "2036885342134173915"
    ],
    "profile_bio": {
      "description": "Democratizing AI research, education, and technologies. New AI learning portal: https://t.co/LRnpZN7L4c",
      "entities": {
        "description": {
          "hashtags": [],
          "symbols": [],
          "urls": [
            {
              "display_url": "academy.dair.ai",
              "expanded_url": "https://academy.dair.ai/",
              "indices": [
                80,
                103
              ],
              "url": "https://t.co/LRnpZN7L4c"
            }
          ],
          "user_mentions": []
        },
        "url": {
          "urls": [
            {
              "display_url": "dair.ai",
              "expanded_url": "https://www.dair.ai/",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/lkqPZtMU5s"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "display_url": "pic.twitter.com/20kcScA7sN",
        "expanded_url": "https://twitter.com/dair_ai/status/2037548967366738131/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {
            "faces": [
              {
                "h": 84,
                "w": 84,
                "x": 167,
                "y": 551
              },
              {
                "h": 418,
                "w": 418,
                "x": 670,
                "y": 352
              }
            ]
          },
          "orig": {
            "faces": [
              {
                "h": 84,
                "w": 84,
                "x": 167,
                "y": 551
              },
              {
                "h": 418,
                "w": 418,
                "x": 670,
                "y": 352
              }
            ]
          }
        },
        "id_str": "2037548964246077440",
        "indices": [
          278,
          301
        ],
        "media_key": "3_2037548964246077440",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARxG0/zSWkAACgACHEbT/YxbwNMAAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABHEbT/NJaQAAKAAIcRtP9jFvA0wAA",
            "media_key": "3_2037548964246077440"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/HEbT_NJaQAAzBzn.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 711,
              "w": 1270,
              "x": 0,
              "y": 0
            },
            {
              "h": 1270,
              "w": 1270,
              "x": 0,
              "y": 0
            },
            {
              "h": 1448,
              "w": 1270,
              "x": 0,
              "y": 0
            },
            {
              "h": 1570,
              "w": 785,
              "x": 0,
              "y": 0
            },
            {
              "h": 1570,
              "w": 1270,
              "x": 0,
              "y": 0
            }
          ],
          "height": 1570,
          "width": 1270
        },
        "sizes": {
          "large": {
            "h": 1570,
            "w": 1270
          }
        },
        "type": "photo",
        "url": "https://t.co/20kcScA7sN"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {
    "hashtags": [],
    "symbols": [],
    "urls": [
      {
        "display_url": "arxiv.org/abs/2603.15255",
        "expanded_url": "https://arxiv.org/abs/2603.15255",
        "indices": [
          913,
          936
        ],
        "url": "https://t.co/8Zn41OBIra"
      },
      {
        "display_url": "academy.dair.ai",
        "expanded_url": "https://academy.dair.ai/",
        "indices": [
          989,
          1012
        ],
        "url": "https://t.co/LRnpZN7L4c"
      }
    ],
    "user_mentions": []
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "article": null
}