{
  "api_version": "1",
  "updated": "2026-11-05",
  "latest_event": "2026-11-04",
  "rights": "© Frontier Log, all rights reserved.",
  "story": {
    "slug": "meta-muse-rise",
    "title": "Meta's Muse Models Climbed to the Top of DeepSWE in Six Months",
    "description": "Meta's Muse models went from 10.0% to 79.0% on DeepSWE v1.1 between April and October 2026, passing every other lab's reported score on Sep 2.",
    "url": "https://frontierlog.ai/stories/meta-muse-rise",
    "date": "2026-11-04",
    "image": "https://frontierlog.ai/og/story-meta-muse-rise.png",
    "event": "2026-10-12-muse-wonder",
    "tags": [
      "model-releases",
      "agents"
    ],
    "steps": [
      {
        "event": "2026-08-05-muse-spark-1-2",
        "date": "2026-08-05",
        "provider": "meta",
        "label": "Muse Code",
        "note": "Terminal coding agent, launched in beta",
        "key": false
      },
      {
        "event": "2026-09-08-muse-agent",
        "date": "2026-09-08",
        "provider": "meta",
        "label": "Muse agent",
        "note": "Personal AI agent in the Muse app and WhatsApp",
        "key": true
      }
    ],
    "tables": [
      {
        "title": "DeepSWE v1.1, Meta's Muse models",
        "unit": "%",
        "rows": [
          {
            "event": "2026-04-08-muse-spark",
            "date": "2026-04-08",
            "provider": "meta",
            "value": 10,
            "printed": "10.0",
            "source": "https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/",
            "where": "Muse Spark 1.1 benchmark table (image), row 'DeepSWE 1.1', column 'Muse Spark'"
          },
          {
            "event": "2026-07-09-muse-spark-1-1",
            "date": "2026-07-09",
            "provider": "meta",
            "value": 53.3,
            "printed": "53.3",
            "source": "https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/",
            "where": "benchmark table (image), row 'DeepSWE 1.1', column 'Muse Spark 1.1'"
          },
          {
            "event": "2026-08-05-muse-spark-1-2",
            "date": "2026-08-05",
            "provider": "meta",
            "value": 59.3,
            "printed": "59.3%",
            "source": "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2",
            "where": "bar chart 'DeepSWE 1.1' (image, Coding tab), value label above the bar 'Muse Spark 1.2, Muse Code'"
          },
          {
            "event": "2026-09-02-muse-spark-1-3",
            "date": "2026-09-02",
            "provider": "meta",
            "value": 75.4,
            "printed": "75.4",
            "source": "https://research.meta.ai/blog/introducing-muse-spark-1-3",
            "where": "scorecard (image), row 'DeepSWE v1.1', column 'Muse Spark 1.3 (max)'"
          },
          {
            "event": "2026-10-12-muse-wonder",
            "date": "2026-10-12",
            "provider": "meta",
            "value": 78.8,
            "printed": "78.8",
            "source": "https://research.meta.ai/blog/muse-wonder",
            "where": "evaluation table (image), row 'DeepSWE 1.1', column 'Muse Wonder (max)'"
          },
          {
            "event": "2026-10-12-muse-wonder",
            "date": "2026-10-12",
            "provider": "meta",
            "value": 79,
            "printed": "79.0",
            "source": "https://research.meta.ai/blog/muse-wonder",
            "where": "evaluation table (image), row 'DeepSWE 1.1', column 'Muse Wonder Full (max)'"
          }
        ]
      },
      {
        "title": "DeepSWE v1.1, other labs' highest score to date",
        "unit": "%",
        "rows": [
          {
            "event": "2026-04-07-glm-5-1",
            "date": "2026-04-07",
            "provider": "zai",
            "value": 18,
            "printed": "18.0",
            "source": "https://z.ai/blog/glm-5.2",
            "where": "'Full Benchmark Table', row 'DeepSWE', column 'GLM-5.1'"
          },
          {
            "event": "2026-04-23-gpt-5-5",
            "date": "2026-04-23",
            "provider": "openai",
            "value": 67,
            "printed": "67%",
            "source": "https://openai.com/index/gpt-5-6",
            "where": "coding eval table (columns GPT‑5.6 Sol, Sol Ultra, Terra, Luna, GPT‑5.5, …), row 'DeepSWE v1.1', column 'GPT‑5.5'"
          },
          {
            "event": "2026-06-09-claude-fable-5-mythos-5",
            "date": "2026-06-09",
            "provider": "anthropic",
            "value": 69.7,
            "printed": "69.7%",
            "source": "https://www.anthropic.com/news/claude-opus-5",
            "where": "benchmark table (image), row 'Agentic coding, DeepSWE v1.1', column 'Fable 5'"
          },
          {
            "event": "2026-07-09-gpt-5-6",
            "date": "2026-07-09",
            "provider": "openai",
            "value": 72.7,
            "printed": "72.7%",
            "source": "https://openai.com/index/gpt-5-6",
            "where": "coding eval table (columns GPT‑5.6 Sol, Sol Ultra, Terra, Luna, GPT‑5.5, …), row 'DeepSWE v1.1', column 'GPT‑5.6 Sol'"
          },
          {
            "event": "2026-09-02-gemini-3-8-flash-and-3-8-flash",
            "date": "2026-09-02",
            "provider": "google",
            "value": 73.7,
            "printed": "73.7%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/",
            "where": "benchmark table (image), row 'DeepSWE v1.1', column 'Gemini 3.8 Flash'"
          },
          {
            "event": "2026-09-03-gpt-6-astra",
            "date": "2026-09-03",
            "provider": "openai",
            "value": 74.1,
            "printed": "74.1%",
            "source": "https://openai.com/index/gpt-6-astra",
            "where": "Coding table, row 'DeepSWE v1.1', column 'GPT‑6 Astra'"
          },
          {
            "event": "2026-09-10-deepseek-v4-1-flash",
            "date": "2026-09-10",
            "provider": "deepseek",
            "value": 74.2,
            "printed": "74.2",
            "source": "https://api-docs.deepseek.com/news/news260910/",
            "where": "benchmark table image (second image), row 'DeepSWE v1.1', column 'DeepSeek V4.1-Flash' (the first chart image, v4.1_260910_benchmark.png, prints the same value label)"
          },
          {
            "event": "2026-09-30-gemini-4-argon",
            "date": "2026-09-30",
            "provider": "google",
            "value": 77.9,
            "printed": "77.9%",
            "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/",
            "where": "benchmark table (image), row 'DeepSWE v1.1', column 'Gemini 4 Argon'"
          }
        ]
      },
      {
        "title": "DeepSWE v1.1, Meta's Muse Spark 1.3 comparison table",
        "unit": "%",
        "rows": [
          {
            "event": "2026-09-02-muse-spark-1-3",
            "date": "2026-09-02",
            "provider": "meta",
            "value": 75.4,
            "printed": "75.4",
            "source": "https://research.meta.ai/blog/introducing-muse-spark-1-3",
            "where": "scorecard (image), row 'DeepSWE v1.1', column 'Muse Spark 1.3 (max)'"
          },
          {
            "event": "2026-08-05-muse-spark-1-2",
            "date": "2026-08-05",
            "provider": "meta",
            "value": 55,
            "printed": "55.0",
            "source": "https://research.meta.ai/blog/introducing-muse-spark-1-3",
            "where": "Muse Spark 1.3 scorecard (image), row 'DeepSWE v1.1', column 'Muse Spark 1.2 (xhigh)'"
          },
          {
            "event": "2026-07-24-claude-opus-5",
            "date": "2026-07-24",
            "provider": "anthropic",
            "value": 74.0,
            "printed": null,
            "source": "https://research.meta.ai/blog/introducing-muse-spark-1-3",
            "where": "scorecard (image), row 'DeepSWE v1.1', column 'Opus 5 (max)'"
          },
          {
            "event": "2026-07-09-gpt-5-6",
            "date": "2026-07-09",
            "provider": "openai",
            "value": 73.0,
            "printed": null,
            "source": "https://research.meta.ai/blog/introducing-muse-spark-1-3",
            "where": "scorecard (image), row 'DeepSWE v1.1', column 'GPT 5.6 Sol (max)'"
          }
        ]
      }
    ],
    "cites": [
      {
        "url": "https://deepswe.datacurve.ai/",
        "by": "Datacurve",
        "medium": "DeepSWE leaderboard",
        "date": "2026-09-22"
      }
    ],
    "sources": [
      "https://ai.meta.com/blog/introducing-muse-spark-msl/",
      "https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/",
      "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2",
      "https://research.meta.ai/blog/introducing-muse-spark-1-3",
      "https://research.meta.ai/blog/muse-wonder",
      "https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/",
      "https://z.ai/blog/glm-5.2",
      "https://www.anthropic.com/news/claude-opus-5",
      "https://openai.com/index/gpt-5-6",
      "https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/",
      "https://openai.com/index/gpt-6-astra",
      "https://api-docs.deepseek.com/news/news260910/",
      "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/",
      "https://deepswe.datacurve.ai/"
    ],
    "body": [
      {
        "figure": "main"
      },
      {
        "p": "Meta's Muse models went from 10.0% to 79.0% on [DeepSWE v1.1](https://frontierlog.ai/benchmark/deepswe-1-1), a long-horizon coding benchmark, in the six months after Meta Superintelligence Labs launched [Muse Spark](https://ai.meta.com/blog/introducing-muse-spark-msl/), the first of its Muse models, on Apr 8, 2026. On Sep 2, [Muse Spark 1.3](https://frontierlog.ai/event/2026-09-02-muse-spark-1-3) scored above every score another lab had reported for its own model. Between those launches, Meta released two agents: Muse Code for coding and Muse, a personal agent."
      },
      {
        "h": "From 10% to 59% in four months"
      },
      {
        "p": "Meta reported Muse Spark's 10.0% in [its Muse Spark 1.1 post on Jul 9](https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/), where Muse Spark 1.1 scored 53.3%. OpenAI launched [GPT-5.6 Sol](https://frontierlog.ai/model/gpt-5-6-sol) the same day and [reported 72.7%](https://openai.com/index/gpt-5-6), above every other lab's model launched by then."
      },
      {
        "p": "On Aug 5, Meta launched [Muse Spark 1.2 and Muse Code](https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2), a terminal coding agent in beta. Muse Spark 1.2 scored 59.3% running in Muse Code. Meta [took Muse Code out of beta](https://frontierlog.ai/event/2026-08-31-muse-code-out-of-beta) on Aug 31."
      },
      {
        "h": "Ahead of the other labs on Sep 2"
      },
      {
        "p": "[Muse Spark 1.3](https://research.meta.ai/blog/introducing-muse-spark-1-3) scored 75.4% at max effort on Sep 2. Google reported 73.7% for [Gemini 3.8 Flash](https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/) the same day. In its own comparison table, Meta put GPT-5.6 Sol at 73.0% and [Claude Opus 5](https://frontierlog.ai/model/claude-opus-5) at 74.0%, both at max effort."
      },
      {
        "p": "Six days later, on Sep 8, Meta launched [Muse](https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/), \"a secure, private personal AI agent\" in the Muse app and WhatsApp, which Meta said is powered by Muse Spark. It began rolling out in the US."
      },
      {
        "p": "Google's [Gemini 4 Argon](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/) passed Muse Spark 1.3's 75.4% on Sep 30, with 77.9%."
      },
      {
        "h": "Muse Wonder Full at 79.0%"
      },
      {
        "p": "On Oct 12, Meta launched [Muse Wonder and Muse Wonder Full](https://research.meta.ai/blog/muse-wonder), which scored 78.8% and 79.0% at max effort. As of Nov 4, no other lab has reported a higher DeepSWE v1.1 score for its own model in the posts Frontier Log records."
      },
      {
        "h": "How the scores compare"
      },
      {
        "p": "Each score is the lab's own report for its own model, with its own harness and effort setting. Meta reported Muse Spark 1.2 at 59.3% in Muse Code at its launch, for example, and at 55.0% at xhigh effort, with no harness named, in its later [Muse Spark 1.3 post](https://research.meta.ai/blog/introducing-muse-spark-1-3). Some earlier scores were reported after their models launched: Muse Spark's in Meta's July post, [GLM-5.1](https://frontierlog.ai/model/glm-5-1)'s 18.0% in [Z.ai's GLM-5.2 post](https://z.ai/blog/glm-5.2) on Jun 16, [GPT-5.5](https://frontierlog.ai/model/gpt-5-5)'s 67% in OpenAI's [GPT-5.6 post](https://openai.com/index/gpt-5-6) on Jul 9, and [Claude Fable 5](https://frontierlog.ai/model/claude-fable-5)'s 69.7% in [Anthropic's Claude Opus 5 post](https://www.anthropic.com/news/claude-opus-5) on Jul 24."
      },
      {
        "h": "Not yet on Datacurve's leaderboard"
      },
      {
        "p": "Datacurve, which makes DeepSWE, runs its own [DeepSWE leaderboard](https://deepswe.datacurve.ai/). As of Nov 4, the leaderboard was last updated on Sep 22. It lists Muse Spark 1.2 at 55% at xhigh effort, and no Muse Spark 1.3, Muse Wonder or Muse Wonder Full. [GPT-6 Astra](https://frontierlog.ai/model/gpt-6-astra), Gemini 3.8 Flash and Claude Opus 5 lead there, at 74% each."
      }
    ],
    "events": [
      "2026-04-08-muse-spark",
      "2026-06-16-glm-5-2",
      "2026-07-09-gpt-5-6",
      "2026-07-09-muse-spark-1-1",
      "2026-07-24-claude-opus-5",
      "2026-08-05-muse-spark-1-2",
      "2026-08-31-muse-code-out-of-beta",
      "2026-09-02-gemini-3-8-flash-and-3-8-flash",
      "2026-09-02-muse-spark-1-3",
      "2026-09-03-gpt-6-astra",
      "2026-09-08-muse-agent",
      "2026-09-10-deepseek-v4-1-flash",
      "2026-09-30-gemini-4-argon",
      "2026-10-12-muse-wonder"
    ],
    "stories": [],
    "models": [
      "muse-spark",
      "muse-spark-1-1",
      "muse-spark-1-2",
      "muse-spark-1-3",
      "muse-wonder",
      "muse-wonder-full",
      "glm-5-1",
      "gpt-5-5",
      "claude-fable-5",
      "gpt-5-6-sol",
      "gemini-3-8-flash",
      "gpt-6-astra",
      "deepseek-v4-1-flash",
      "gemini-4-argon",
      "claude-opus-5"
    ]
  }
}