{
  "results": [
    {
      "id": "spanish_vs_english_cost",
      "expected": "comparar consumo aproximado de tokens entre idiomas",
      "must_be_similar": false,
      "texts": [
        {
          "text": "machine learning is useful",
          "tokens": [
            "machine",
            "learning",
            "is",
            "useful"
          ],
          "token_ids": [
            23723,
            3134,
            35838,
            41931
          ],
          "token_count": 4,
          "estimated_cost_usd": 8e-08,
          "embedding_preview": [
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.408,
            0.0
          ]
        },
        {
          "text": "el aprendizaje automático resulta útil",
          "tokens": [
            "el",
            "aprendizaje",
            "automático",
            "resulta",
            "útil"
          ],
          "token_ids": [
            8368,
            3474,
            9725,
            370,
            1710
          ],
          "token_count": 5,
          "estimated_cost_usd": 1e-07,
          "embedding_preview": [
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0
          ]
        }
      ],
      "cosine_similarity": 0.7715,
      "passes_similarity_threshold": true,
      "token_budget_warning": false
    },
    {
      "id": "semantic_pair",
      "expected": "deberían quedar cerca aunque compartan pocas palabras",
      "must_be_similar": true,
      "texts": [
        {
          "text": "gato doméstico",
          "tokens": [
            "gato",
            "doméstico"
          ],
          "token_ids": [
            28189,
            36126
          ],
          "token_count": 2,
          "estimated_cost_usd": 4e-08,
          "embedding_preview": [
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0
          ]
        },
        {
          "text": "felino de casa",
          "tokens": [
            "felino",
            "de",
            "casa"
          ],
          "token_ids": [
            44852,
            1675,
            39475
          ],
          "token_count": 3,
          "estimated_cost_usd": 6e-08,
          "embedding_preview": [
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0,
            0.0
          ]
        }
      ],
      "cosine_similarity": 0.8165,
      "passes_similarity_threshold": true,
      "token_budget_warning": false
    },
    {
      "id": "code_pair",
      "expected": "misma intención de código en sintaxis distinta",
      "must_be_similar": true,
      "texts": [
        {
          "text": "function getUser(id) { return db.user(id) }",
          "tokens": [
            "function",
            "getuser",
            "(",
            "id",
            ")",
            "{",
            "return",
            "db",
            ".",
            "user",
            "(",
            "id",
            ")",
            "}"
          ],
          "token_ids": [
            11038,
            10363,
            35641,
            33644,
            347,
            21677,
            20319,
            48780,
            5336,
            32255,
            35641,
            33644,
            347,
            30178
          ],
          "token_count": 14,
          "estimated_cost_usd": 2.8e-07,
          "embedding_preview": [
            0.0,
            0.0,
            -0.236,
            0.0,
            0.236,
            -0.471,
            0.236,
            0.0
          ]
        },
        {
          "text": "def get_user(id): return db.user(id)",
          "tokens": [
            "def",
            "get_user",
            "(",
            "id",
            ")",
            ":",
            "return",
            "db",
            ".",
            "user",
            "(",
            "id",
            ")"
          ],
          "token_ids": [
            27379,
            33042,
            35641,
            33644,
            347,
            21302,
            20319,
            48780,
            5336,
            32255,
            35641,
            33644,
            347
          ],
          "token_count": 13,
          "estimated_cost_usd": 2.6e-07,
          "embedding_preview": [
            0.0,
            0.0,
            0.0,
            0.0,
            0.258,
            -0.516,
            0.258,
            0.258
          ]
        }
      ],
      "cosine_similarity": 0.9129,
      "passes_similarity_threshold": true,
      "token_budget_warning": true
    },
    {
      "id": "long_word",
      "expected": "ver subtokens y parentesco morfológico",
      "must_be_similar": true,
      "texts": [
        {
          "text": "desarrolladores responsables",
          "tokens": [
            "desa",
            "rroll",
            "adore",
            "s",
            "responsables"
          ],
          "token_ids": [
            20798,
            2513,
            43312,
            17434,
            25299
          ],
          "token_count": 5,
          "estimated_cost_usd": 1e-07,
          "embedding_preview": [
            0.0,
            0.0,
            0.0,
            -0.302,
            0.302,
            0.0,
            0.0,
            0.0
          ]
        },
        {
          "text": "desarrollo responsable",
          "tokens": [
            "desarrollo",
            "responsable"
          ],
          "token_ids": [
            29925,
            32326
          ],
          "token_count": 2,
          "estimated_cost_usd": 4e-08,
          "embedding_preview": [
            0.0,
            0.0,
            0.0,
            -0.707,
            0.0,
            0.0,
            0.0,
            0.0
          ]
        }
      ],
      "cosine_similarity": 0.8528,
      "passes_similarity_threshold": true,
      "token_budget_warning": false
    }
  ],
  "invalid": []
}