{
  "api_version": "1",
  "updated": "2026-11-05",
  "latest_event": "2026-11-04",
  "rights": "© Frontier Log, all rights reserved.",
  "model": {
    "slug": "laguna-xs-2",
    "name": "Laguna XS.2",
    "lab": "poolside",
    "date": "2026-04-28",
    "source": "https://poolside.ai/blog/introducing-laguna-xs2-m1",
    "source_title": "Introducing Laguna XS.2 and Laguna M.1",
    "flagship": false,
    "event": "https://frontierlog.ai/event/2026-04-28-laguna-xs-2-and-m-1",
    "releases": [],
    "variant_of": null,
    "made_by": "Poolside",
    "url": "https://frontierlog.ai/model/laguna-xs-2",
    "results": [
      {
        "id": "laguna-xs-2:swe-bench-verified:default:1",
        "model": "laguna-xs-2",
        "model_name": "Laguna XS.2",
        "model_date": "2026-04-28",
        "event": "https://frontierlog.ai/event/2026-04-28-laguna-xs-2-and-m-1",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 68.2,
        "unit": "%",
        "printed": "68.2",
        "reported_by": "poolside",
        "run_by": null,
        "label": "Laguna XS.2 (33B-A3B)",
        "source": "https://poolside.ai/blog/introducing-laguna-xs2-m1",
        "source_title": "Introducing Laguna XS.2 and Laguna M.1",
        "source_date": "2026-04-28",
        "where": "the Laguna XS.2 comparison chart ('(Extra) Small model, big story' section), benchmark panel 'SWE-bench Verified' (Resolved tasks on SWE-bench Verified), value in the column 'Laguna XS.2 (33B-A3B)' (first column; chart rendered by script, read the rendered page)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "poolside harness on Harbor; pass@1 averaged over 4 runs; some task images patched",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-11-05",
        "corroborated_by": [
          {
            "url": "https://poolside.ai/blog/laguna-a-deeper-dive",
            "name": "Laguna XS.2 and Laguna M.1 card",
            "title": "Laguna XS.2 and M.1: A Deeper Dive",
            "lab": "poolside",
            "date": "2026-04-28",
            "printed": "68.2",
            "where": "SWE-bench Verified bar chart, aria-label 'Laguna XS.2 (33B-A3B): 68.2, highlighted', and the HTML table row 'SWE-bench Verified', Laguna XS.2 (33B-A3B) column (Benchmarks as of 28 April 2026)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/9d496a87332363b8.png"
      },
      {
        "id": "laguna-xs-2:swe-bench-pro:default:1",
        "model": "laguna-xs-2",
        "model_name": "Laguna XS.2",
        "model_date": "2026-04-28",
        "event": "https://frontierlog.ai/event/2026-04-28-laguna-xs-2-and-m-1",
        "benchmark": "swe-bench-pro",
        "benchmark_name": "SWE-bench Pro",
        "setting": "default",
        "setting_name": "",
        "score": 44.5,
        "unit": "%",
        "printed": "44.5",
        "reported_by": "poolside",
        "run_by": null,
        "label": "Laguna XS.2 (33B-A3B)",
        "source": "https://poolside.ai/blog/introducing-laguna-xs2-m1",
        "source_title": "Introducing Laguna XS.2 and Laguna M.1",
        "source_date": "2026-04-28",
        "where": "the Laguna XS.2 comparison chart ('(Extra) Small model, big story' section), benchmark panel 'SWE-Bench Pro' (Resolved tasks on SWE-Bench Pro), value in the column 'Laguna XS.2 (33B-A3B)' (first column; chart rendered by script, read the rendered page)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "poolside harness on Harbor; pass@1 averaged over 3 runs; some task images patched",
        "caveat": "The source doesn't say whether this is the public or the commercial set",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-11-05",
        "corroborated_by": [
          {
            "url": "https://poolside.ai/blog/laguna-a-deeper-dive",
            "name": "Laguna XS.2 and Laguna M.1 card",
            "title": "Laguna XS.2 and M.1: A Deeper Dive",
            "lab": "poolside",
            "date": "2026-04-28",
            "printed": "44.5",
            "where": "SWE-Bench Pro bar chart, aria-label 'Laguna XS.2 (33B-A3B): 44.5, highlighted', and the HTML table row 'SWE-Bench Pro', Laguna XS.2 (33B-A3B) column (Benchmarks as of 28 April 2026)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/9d496a87332363b8.png"
      },
      {
        "id": "laguna-xs-2:terminal-bench-2-0:default:1",
        "model": "laguna-xs-2",
        "model_name": "Laguna XS.2",
        "model_date": "2026-04-28",
        "event": "https://frontierlog.ai/event/2026-04-28-laguna-xs-2-and-m-1",
        "benchmark": "terminal-bench-2-0",
        "benchmark_name": "Terminal-Bench 2.0",
        "setting": "default",
        "setting_name": "",
        "score": 30.1,
        "unit": "%",
        "printed": "30.1",
        "reported_by": "poolside",
        "run_by": null,
        "label": "Laguna XS.2 (33B-A3B)",
        "source": "https://poolside.ai/blog/introducing-laguna-xs2-m1",
        "source_title": "Introducing Laguna XS.2 and Laguna M.1",
        "source_date": "2026-04-28",
        "where": "the Laguna XS.2 comparison chart ('(Extra) Small model, big story' section), benchmark panel 'Terminal-Bench 2.0' (Resolved tasks on Terminal-Bench 2.0), value in the column 'Laguna XS.2 (33B-A3B)' (first column; chart rendered by script, read the rendered page)",
        "own": true,
        "primary": true,
        "parallel": false,
        "notes": "poolside harness on Harbor; pass@1 averaged over 5 runs; 48GB RAM/32 CPUs; some task images patched",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-11-05",
        "corroborated_by": [
          {
            "url": "https://poolside.ai/blog/laguna-a-deeper-dive",
            "name": "Laguna XS.2 and Laguna M.1 card",
            "title": "Laguna XS.2 and M.1: A Deeper Dive",
            "lab": "poolside",
            "date": "2026-04-28",
            "printed": "30.1",
            "where": "Terminal-Bench 2.0 bar chart, aria-label 'Laguna XS.2 (33B-A3B): 30.1, highlighted', and the HTML table row 'Terminal-Bench 2.0', Laguna XS.2 (33B-A3B) column (Benchmarks as of 28 April 2026)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/9d496a87332363b8.png"
      },
      {
        "id": "laguna-xs-2:swe-bench-verified:default:2",
        "model": "laguna-xs-2",
        "model_name": "Laguna XS.2",
        "model_date": "2026-04-28",
        "event": "https://frontierlog.ai/event/2026-04-28-laguna-xs-2-and-m-1",
        "benchmark": "swe-bench-verified",
        "benchmark_name": "SWE-bench Verified",
        "setting": "default",
        "setting_name": "",
        "score": 69.9,
        "unit": "%",
        "printed": "69.9",
        "reported_by": "poolside",
        "run_by": null,
        "label": "Laguna XS.2 (33B-A3B)",
        "source": "https://poolside.ai/blog/long-context-update-laguna-xs-2-and-m-1",
        "source_title": "Long context update: Laguna XS.2 and M.1",
        "source_date": "2026-05-26",
        "where": "the single comparison chart under 'Laguna M.1 remains our most capable model', benchmark panel 'SWE-bench Verified' (Resolved tasks on SWE-bench Verified), value in the column 'Laguna XS.2 (33B-A3B)' (chart rendered by script, read the rendered page)",
        "own": false,
        "primary": false,
        "parallel": false,
        "notes": "poolside harness on Harbor; pass@1 averaged over 4 runs; some task images patched",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-11-05",
        "corroborated_by": [
          {
            "url": "https://huggingface.co/poolside/Laguna-XS.2",
            "name": "Laguna XS.2 and Laguna M.1 card",
            "title": "Laguna XS.2 model card",
            "lab": "poolside",
            "date": "2026-04-28",
            "printed": "69.9%",
            "where": "Benchmark results table, row '**Laguna XS.2**', column 'SWE-bench Verified' (README.md)"
          },
          {
            "url": "https://poolside.ai/blog/introducing-laguna-xs-2-1",
            "name": "Laguna XS 2.1",
            "title": "Introducing Laguna XS 2.1",
            "lab": "poolside",
            "date": "2026-07-02",
            "printed": "69.9",
            "where": "XS 2.1 vs XS.2 bar chart, SWE-bench Verified, aria-label 'Laguna XS.2 (33B-A3B): 69.9' (comparison bar, Benchmarks as of 2 July 2026)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/c192588e44b50c33.png"
      },
      {
        "id": "laguna-xs-2:swe-bench-pro:default:2",
        "model": "laguna-xs-2",
        "model_name": "Laguna XS.2",
        "model_date": "2026-04-28",
        "event": "https://frontierlog.ai/event/2026-04-28-laguna-xs-2-and-m-1",
        "benchmark": "swe-bench-pro",
        "benchmark_name": "SWE-bench Pro",
        "setting": "default",
        "setting_name": "",
        "score": 46.3,
        "unit": "%",
        "printed": "46.3",
        "reported_by": "poolside",
        "run_by": null,
        "label": "Laguna XS.2 (33B-A3B)",
        "source": "https://poolside.ai/blog/long-context-update-laguna-xs-2-and-m-1",
        "source_title": "Long context update: Laguna XS.2 and M.1",
        "source_date": "2026-05-26",
        "where": "the single comparison chart under 'Laguna M.1 remains our most capable model', benchmark panel 'SWE-Bench Pro' (Resolved tasks on SWE-Bench Pro), value in the column 'Laguna XS.2 (33B-A3B)' (chart rendered by script, read the rendered page)",
        "own": false,
        "primary": false,
        "parallel": false,
        "notes": "poolside harness on Harbor; pass@1 averaged over 3 runs; some task images patched",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-11-05",
        "corroborated_by": [
          {
            "url": "https://huggingface.co/poolside/Laguna-XS.2",
            "name": "Laguna XS.2 and Laguna M.1 card",
            "title": "Laguna XS.2 model card",
            "lab": "poolside",
            "date": "2026-04-28",
            "printed": "46.3%",
            "where": "Benchmark results table, row '**Laguna XS.2**', column 'SWE-bench Pro (Public Dataset)' (README.md)"
          },
          {
            "url": "https://poolside.ai/blog/introducing-laguna-xs-2-1",
            "name": "Laguna XS 2.1",
            "title": "Introducing Laguna XS 2.1",
            "lab": "poolside",
            "date": "2026-07-02",
            "printed": "46.3",
            "where": "XS 2.1 vs XS.2 bar chart, SWE-Bench Pro, aria-label 'Laguna XS.2 (33B-A3B): 46.3' (comparison bar, Benchmarks as of 2 July 2026)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/c192588e44b50c33.png"
      },
      {
        "id": "laguna-xs-2:terminal-bench-2-0:default:2",
        "model": "laguna-xs-2",
        "model_name": "Laguna XS.2",
        "model_date": "2026-04-28",
        "event": "https://frontierlog.ai/event/2026-04-28-laguna-xs-2-and-m-1",
        "benchmark": "terminal-bench-2-0",
        "benchmark_name": "Terminal-Bench 2.0",
        "setting": "default",
        "setting_name": "",
        "score": 35.7,
        "unit": "%",
        "printed": "35.7",
        "reported_by": "poolside",
        "run_by": null,
        "label": "Laguna XS.2 (33B-A3B)",
        "source": "https://poolside.ai/blog/long-context-update-laguna-xs-2-and-m-1",
        "source_title": "Long context update: Laguna XS.2 and M.1",
        "source_date": "2026-05-26",
        "where": "the single comparison chart under 'Laguna M.1 remains our most capable model', benchmark panel 'Terminal-Bench 2.0' (Resolved tasks on Terminal-Bench 2.0), value in the column 'Laguna XS.2 (33B-A3B)' (chart rendered by script, read the rendered page)",
        "own": false,
        "primary": false,
        "parallel": false,
        "notes": "poolside harness on Harbor; pass@1 averaged over 5 runs; 48GB RAM/32 CPUs; some task images patched",
        "caveat": "",
        "human_verified": false,
        "human_verified_on": null,
        "check": "corroborated",
        "corroborated": true,
        "corroborated_on": "2026-11-05",
        "corroborated_by": [
          {
            "url": "https://huggingface.co/poolside/Laguna-XS.2",
            "name": "Laguna XS.2 and Laguna M.1 card",
            "title": "Laguna XS.2 model card",
            "lab": "poolside",
            "date": "2026-04-28",
            "printed": "35.7%",
            "where": "Benchmark results table, row '**Laguna XS.2**', column 'Terminal-Bench 2.0' (README.md)"
          },
          {
            "url": "https://poolside.ai/blog/introducing-laguna-xs-2-1",
            "name": "Laguna XS 2.1",
            "title": "Introducing Laguna XS 2.1",
            "lab": "poolside",
            "date": "2026-07-02",
            "printed": "35.7",
            "where": "XS 2.1 vs XS.2 bar chart, Terminal-Bench 2.0, aria-label 'Laguna XS.2 (33B-A3B): 35.7' (comparison bar, Benchmarks as of 2 July 2026)"
          }
        ],
        "source_capture": "https://frontierlog.ai/captures/c192588e44b50c33.png"
      }
    ],
    "price": [],
    "context": {
      "from": "2026-05-26",
      "listed": false,
      "window_tokens": 256000,
      "input_tokens": null,
      "output_tokens": null,
      "tiers": [],
      "printed": {
        "window": "256K",
        "input": null,
        "output": null
      },
      "notes": "",
      "sources": [
        "https://poolside.ai/blog/long-context-update-laguna-xs-2-and-m-1",
        "https://huggingface.co/poolside/Laguna-XS.2"
      ],
      "check": "corroborated"
    },
    "prices": [],
    "context_history": [
      {
        "from": "2026-05-26",
        "listed": false,
        "window_tokens": 256000,
        "input_tokens": null,
        "output_tokens": null,
        "tiers": [],
        "printed": {
          "window": "256K",
          "input": null,
          "output": null
        },
        "notes": "",
        "sources": [
          "https://poolside.ai/blog/long-context-update-laguna-xs-2-and-m-1",
          "https://huggingface.co/poolside/Laguna-XS.2"
        ],
        "check": "corroborated"
      }
    ],
    "stories": []
  }
}