Analisi III · Exercise research profile

Livelli di R3→R2, sviluppo quadratico, Lagrange ed estremi assoluti (A3-25-P4-E1)

Source record, frozen scoring evidence, configuration-level measurements and retained responses for one benchmark exercise.

01

Observed panel summary

Exercise overview

Descriptive results across 21 model-mode configuration means. They do not establish general model superiority.

Published recordSource statement · frozen rubric · 23 retained responses · current judgmentsThe separate official reference solution is not reproduced.
Mean rubric score93.6Median 99.5 · n=16
Configuration spreadSD 10.564.5–100.0
Rank by lowest observed mean#36/60Overall · #25/33 within Analisi III
Mean latency256.4sEnd-to-end per response
Mean cost$0.116Snapshot estimate
Mean tokens13,665Analytical tokens per response

Configuration means

Observed score distribution

Median 99.5 / 100
AnthropicDeepSeekOpenAIGooglexAIMistral
  1. Claude Fable 5SD not available · single run
    100.0
  2. Claude Opus 5.5SD not available · single run
    100.0
  3. DeepSeek V4 FlashSD not available · single run
    100.0
  4. DeepSeek V4 Flash (no thinking)SD not available · single run
    100.0
  5. DeepSeek V4 ProSD not available · single run
    100.0
  6. Gemini 3.5 FlashSD 0.0 · 3 retained runs
    100.0
  7. GPT-5.6 SolSD not available · single run
    100.0
  8. GPT-6 LunaSD not available · single run
    100.0
  9. GPT-6.1 SolSD not available · single run
    100.0
  10. Grok 4.5SD not available · single run
    100.0
  11. Claude Sonnet 5.5SD not available · single run
    99.5
  12. OpenAI o3SD not available · single run
    97.9
  13. Gemma 4 31B ITSD not available · single run
    96.6
  14. Mistral Small 4SD not available · single run
    96.6
  15. DeepSeek V4.1 FlashSD not available · single run
    95.3
  16. Gemini 3.1 Pro PreviewSD not available · single run
    94.2
  17. Mistral Medium 3.5SD not available · single run
    92.6
  18. OpenAI GPT-4.1SD not available · single run
    81.8
  19. Claude Sonnet 4.5SD not available · single run
    77.9
  20. Grok 4.3SD not available · single run
    69.5
  21. Claude Haiku 4.5SD not available · single run
    64.5

Whiskers show mean ± one within-configuration response SD where three retained runs are available. Single-run cells have no within-configuration variability statistic.

02

Task before results

Source record and frozen rubric

Read the evaluated task before interpreting configuration outcomes. Original Italian records remain authoritative.

Evaluated source record

Exercise statement

A3-25-P4-E1 · Analisi III
The English version is a faithful editorial translation. The Italian source remains authoritative.

ESAME: Analisi 3 DATA: 03/07/2025 ESERCIZIO: 1

TESTO: 1. Nello spazio cartesiano sia ( g,h ) : R3→R2 data da ( g(x,y,z ),h (x,y,z )) = (2xy +z, yz−z2). (a) Provare che l’insieme di livello Γ di (g,h ) passante perP (0, 1,−1) è una curva regolare attorno aP , parametrizzarla inP fino all’ordine quadratico e calcolarne in due modi la retta tangente affine in P . Determinare poi in generale quali insiemi di livello di ( g,h ) sono curve regolari. (b) Trovare i punti stazionari su Γ per la funzione f(x,y,z ) =y−3z; mostrare cheP è tra questi, determinandone la natura. (c) Disegnare l’insiemeB ={(x,y,z ) :z = 1, x2−1≤y≤1}e calcolare gli estremi assoluti di g su B, spiegando perché esistono.

Source transcription errata1 recorded normalization · original preserved+

These notes document evident transcription or OCR artifacts. They do not overwrite the authoritative source record above.

As transcribedR3→R2; yz−z2; x2−1; perP; insiemeBEditorial normalizationℝ³→ℝ²; yz−z²; x²−1; per P; insieme BLost superscripts, number-set styling and spaces.
Original source textExact Italian record · no editorial changes+
ESAME: Analisi 3
DATA: 03/07/2025
ESERCIZIO: 1

TESTO:
1. Nello spazio cartesiano sia ( g,h ) : R3→R2 data da ( g(x,y,z ),h (x,y,z )) = (2xy +z, yz−z2).
(a) Provare che l’insieme di livello Γ di (g,h ) passante perP (0, 1,−1) è una curva regolare attorno
aP , parametrizzarla inP fino all’ordine quadratico e calcolarne in due modi la retta tangente
affine in P . Determinare poi in generale quali insiemi di livello di ( g,h ) sono curve regolari.
(b) Trovare i punti stazionari su Γ per la funzione f(x,y,z ) =y−3z; mostrare cheP è tra questi,
determinandone la natura.
(c) Disegnare l’insiemeB ={(x,y,z ) :z = 1, x2−1≤y≤1}e calcolare gli estremi assoluti di g
su B, spiegando perché esistono.
TopicLivelli di R3→R2, sviluppo quadratico, Lagrange ed estremi assoluti
Syllabus areaCurve di livello ed estremi vincolati
Task typemixed
Atomic criteria23
Subproblem a · 43.2%

Regolarità, sviluppo, tangente in due modi e livelli generali.

9 atomic criteria · T1 · T2 · T3 · T4 · T5 · T6 · T7

Inspect scoring criteria +
  • Identifica Γ e la regolarità locale.

    T1 · 6 pts · Criterion (a_T1_01)
  • Classifica luogo critico e livelli eccezionali.

    T1 · 4 pts · Criterion (a_T1_02)
  • Imposta implicita, sviluppo e due tangenti.

    T2 · 6 pts · Criterion (a_T2_01)
  • Calcola derivate prime.

    T3 · 5 pts · Criterion (a_T3_01)
  • Calcola derivate seconde.

    T3 · 5 pts · Criterion (a_T3_02)
  • Dà sviluppo, tangente e classificazione.

    T4 · 5 pts · Criterion (a_T4_01)
  • Gestisce rango e località.

    T5 · 4 pts · Criterion (a_T5_01)
  • Controlla sviluppo e tangente.

    T6 · 3 pts · Criterion (a_T6_01)
  • Usa correttamente teorema del livello.

    T7 · 3 pts · Criterion (a_T7_01)
Subproblem b · 23.2%

Punti stazionari e natura di P.

7 atomic criteria · T1 · T2 · T3 · T4 · T5 · T6 · T7

Inspect scoring criteria +
  • Formalizza Lagrange su Γ.

    T1 · 4 pts · Criterion (b_T1_01)
  • Sceglie ricerca e classificazione valide.

    T2 · 4 pts · Criterion (b_T2_01)
  • Risolve e calcola F''.

    T3 · 4 pts · Criterion (b_T3_01)
  • Dà punti e natura.

    T4 · 4 pts · Criterion (b_T4_01)
  • Rispetta vincoli e località.

    T5 · 2 pts · Criterion (b_T5_01)
  • Controlla stazionarietà e segno.

    T6 · 2 pts · Criterion (b_T6_01)
  • Applica Lagrange/test secondo ordine.

    T7 · 2 pts · Criterion (b_T7_01)
Subproblem c · 33.7%

Disegno di B ed estremi assoluti.

7 atomic criteria · T1 · T2 · T3 · T4 · T5 · T6 · T7

Inspect scoring criteria +
  • Descrive B e tutte le parti.

    T1 · 6 pts · Criterion (c_T1_01)
  • Imposta ricerca esaustiva.

    T2 · 5 pts · Criterion (c_T2_01)
  • Calcola candidati e valori.

    T3 · 6 pts · Criterion (c_T3_01)
  • Fornisce esistenza ed estremi.

    T4 · 6 pts · Criterion (c_T4_01)
  • Giustifica compattezza e domini.

    T5 · 4 pts · Criterion (c_T5_01)
  • Confronta i valori.

    T6 · 3 pts · Criterion (c_T6_01)
  • Applica Weierstrass.

    T7 · 2 pts · Criterion (c_T7_01)
Complete frozen rubricOriginal Italian record · exact JSON · SHA-256 24463af1e99e…+
{
  "rubric_version": "1.0",
  "judge_spec_version": "2.0",
  "item_id": "analisi_3_2025_quarto_appello_esercizio_1",
  "syllabus_macroarea": "Curve di livello ed estremi vincolati",
  "topic": "Livelli di R3→R2, sviluppo quadratico, Lagrange ed estremi assoluti",
  "task_type": "mixed",
  "created_without_model_answers": true,
  "frozen": true,
  "reference_status": "possible_issue",
  "reference_notes": "Gold ufficiale. Γ={(g,h)=(-1,-2)}. In P: γ(x)=(x,1-6x+44x²+o(x²),-1-2x+12x²+o(x²)); tangente {2x+z=-1,y-3z=4}. Luogo critico: (-1/4,0,α) e (β,0,0), con livelli potenzialmente singolari (α,-α²), incluso (0,0). I punti stazionari distinti su Γ sono P=(0,1,-1), Q=(1,-1,1); P minimo locale stretto. La reference duplica Q come R: refuso. Su B: minimo 1-2√2 in (-√2,1,1), massimo 1+2√2 in (√2,1,1).",
  "subproblems": [
    {
      "id": "a",
      "weight": 0.43157894736842106,
      "request": "Regolarità, sviluppo, tangente in due modi e livelli generali.",
      "expected_results": [
        "rank J(P)=2",
        "derivate (-6,-2),(88,24)",
        "γ, tangente e livelli eccezionali gold"
      ],
      "required_conditions": [
        "Teorema implicito locale e rango 2."
      ],
      "accepted_equivalences": [
        "Forme equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodi rigorosi compatibili col gold."
      ]
    },
    {
      "id": "b",
      "weight": 0.23157894736842105,
      "request": "Punti stazionari e natura di P.",
      "expected_results": [
        "P,Q",
        "F''(0)=16>0"
      ],
      "required_conditions": [
        "Appartenenza a Γ."
      ],
      "accepted_equivalences": [
        "Forme equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodi rigorosi compatibili col gold."
      ]
    },
    {
      "id": "c",
      "weight": 0.3368421052631579,
      "request": "Disegno di B ed estremi assoluti.",
      "expected_results": [
        "B compatto",
        "candidati completi",
        "estremi gold"
      ],
      "required_conditions": [
        "Analizzare interno e tutto il bordo."
      ],
      "accepted_equivalences": [
        "Forme equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodi rigorosi compatibili col gold."
      ]
    }
  ],
  "atomic_criteria": [
    {
      "criterion_id": "a_T1_01",
      "subproblem_id": "a",
      "primary_dimension": "T1",
      "max_points": 6,
      "description": "Identifica Γ e la regolarità locale.",
      "full_credit_evidence": [
        "Livello (-1,-2), rank J(P)=2."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T1_02",
      "subproblem_id": "a",
      "primary_dimension": "T1",
      "max_points": 4,
      "description": "Classifica luogo critico e livelli eccezionali.",
      "full_credit_evidence": [
        "Famiglie e valori gold."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T2_01",
      "subproblem_id": "a",
      "primary_dimension": "T2",
      "max_points": 6,
      "description": "Imposta implicita, sviluppo e due tangenti.",
      "full_credit_evidence": [
        "Esplicita (y,z)(x) e usa ker J(P)."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T3_01",
      "subproblem_id": "a",
      "primary_dimension": "T3",
      "max_points": 5,
      "description": "Calcola derivate prime.",
      "full_credit_evidence": [
        "y'=-6,z'=-2."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T3_02",
      "subproblem_id": "a",
      "primary_dimension": "T3",
      "max_points": 5,
      "description": "Calcola derivate seconde.",
      "full_credit_evidence": [
        "y''=88,z''=24."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T4_01",
      "subproblem_id": "a",
      "primary_dimension": "T4",
      "max_points": 5,
      "description": "Dà sviluppo, tangente e classificazione.",
      "full_credit_evidence": [
        "Tutti i deliverable gold."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T5_01",
      "subproblem_id": "a",
      "primary_dimension": "T5",
      "max_points": 4,
      "description": "Gestisce rango e località.",
      "full_credit_evidence": [
        "Ipotesi implicite complete."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T6_01",
      "subproblem_id": "a",
      "primary_dimension": "T6",
      "max_points": 3,
      "description": "Controlla sviluppo e tangente.",
      "full_credit_evidence": [
        "J(P)γ'(0)=0."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "a_T7_01",
      "subproblem_id": "a",
      "primary_dimension": "T7",
      "max_points": 3,
      "description": "Usa correttamente teorema del livello.",
      "full_credit_evidence": [
        "Inferenze di rango rigorose."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "b_T1_01",
      "subproblem_id": "b",
      "primary_dimension": "T1",
      "max_points": 4,
      "description": "Formalizza Lagrange su Γ.",
      "full_credit_evidence": [
        "Determinante più vincoli."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "b_T2_01",
      "subproblem_id": "b",
      "primary_dimension": "T2",
      "max_points": 4,
      "description": "Sceglie ricerca e classificazione valide.",
      "full_credit_evidence": [
        "Lagrange e restrizione locale."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "b_T3_01",
      "subproblem_id": "b",
      "primary_dimension": "T3",
      "max_points": 4,
      "description": "Risolve e calcola F''.",
      "full_credit_evidence": [
        "P,Q e 16."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "b_T4_01",
      "subproblem_id": "b",
      "primary_dimension": "T4",
      "max_points": 4,
      "description": "Dà punti e natura.",
      "full_credit_evidence": [
        "P minimo locale stretto."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "b_T5_01",
      "subproblem_id": "b",
      "primary_dimension": "T5",
      "max_points": 2,
      "description": "Rispetta vincoli e località.",
      "full_credit_evidence": [
        "Candidati su Γ."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "b_T6_01",
      "subproblem_id": "b",
      "primary_dimension": "T6",
      "max_points": 2,
      "description": "Controlla stazionarietà e segno.",
      "full_credit_evidence": [
        "F'=0,F''>0."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "b_T7_01",
      "subproblem_id": "b",
      "primary_dimension": "T7",
      "max_points": 2,
      "description": "Applica Lagrange/test secondo ordine.",
      "full_credit_evidence": [
        "Uso rigoroso."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "c_T1_01",
      "subproblem_id": "c",
      "primary_dimension": "T1",
      "max_points": 6,
      "description": "Descrive B e tutte le parti.",
      "full_credit_evidence": [
        "Interno, parabola, segmento, estremi."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "c_T2_01",
      "subproblem_id": "c",
      "primary_dimension": "T2",
      "max_points": 5,
      "description": "Imposta ricerca esaustiva.",
      "full_credit_evidence": [
        "Interno e bordo."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "c_T3_01",
      "subproblem_id": "c",
      "primary_dimension": "T3",
      "max_points": 6,
      "description": "Calcola candidati e valori.",
      "full_credit_evidence": [
        "D,M±,A± gold."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "c_T4_01",
      "subproblem_id": "c",
      "primary_dimension": "T4",
      "max_points": 6,
      "description": "Fornisce esistenza ed estremi.",
      "full_credit_evidence": [
        "Min/max con punti gold."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "c_T5_01",
      "subproblem_id": "c",
      "primary_dimension": "T5",
      "max_points": 4,
      "description": "Giustifica compattezza e domini.",
      "full_credit_evidence": [
        "B compatto, g continua."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "c_T6_01",
      "subproblem_id": "c",
      "primary_dimension": "T6",
      "max_points": 3,
      "description": "Confronta i valori.",
      "full_credit_evidence": [
        "Ordine e appartenenza."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    },
    {
      "criterion_id": "c_T7_01",
      "subproblem_id": "c",
      "primary_dimension": "T7",
      "max_points": 2,
      "description": "Applica Weierstrass.",
      "full_credit_evidence": [
        "Esistenza rigorosa."
      ],
      "partial_credit_rules": [
        {
          "credit_fraction": 0.75,
          "anchor": "Corretto salvo un difetto locale."
        },
        {
          "credit_fraction": 0.5,
          "anchor": "Copertura sostanziale ma incompleta."
        },
        {
          "credit_fraction": 0.25,
          "anchor": "Solo impostazione o elemento utile."
        }
      ],
      "zero_credit_conditions": [
        "Assente o incompatibile col gold."
      ],
      "accepted_equivalences": [
        "Forme matematicamente equivalenti."
      ],
      "accepted_alternative_methods": [
        "Metodo alternativo rigoroso compatibile col gold."
      ],
      "propagated_error_policy": "Penalizzare l'origine una sola volta."
    }
  ],
  "dimension_applicability": {
    "T1": "applicable",
    "T2": "applicable",
    "T3": "applicable",
    "T4": "applicable",
    "T5": "applicable",
    "T6": "applicable",
    "T7": "applicable",
    "T8": "not_applicable"
  },
  "dimension_applicability_notes": {
    "T8": "Non applicabile: nessuna verifica autonoma richiesta."
  },
  "dimension_point_budget": {
    "T1": 20,
    "T2": 15,
    "T3": 20,
    "T4": 15,
    "T5": 10,
    "T6": 8,
    "T7": 7,
    "T8": 0
  },
  "applicable_max_score": 95,
  "normalization_rule": "Normalizzare con 100*raw_score/95.",
  "not_applicable_dimensions": [
    "T8"
  ],
  "pertinent_checks": [
    "J(P)γ'(0)=0.",
    "Luogo rank<2.",
    "Refuso Q/R.",
    "Confronto valori su B."
  ],
  "known_error_severity": [
    {
      "error": "Livelli o sviluppo errati",
      "severity": "major"
    },
    {
      "error": "Punti stazionari/natura errati",
      "severity": "major"
    },
    {
      "error": "Estremi assoluti errati",
      "severity": "major"
    }
  ],
  "rubric_specific_caps": [],
  "human_review_triggers": [
    "Q e R trattati come distinti per il refuso.",
    "Classificazione alternativa incerta."
  ],
  "human_approval": {
    "approved": true,
    "reviewer": "benchmark_owner_via_codex_task",
    "date": "2026-07-18"
  }
}
03

Per-exercise ranking

Configurations

Choose one of four stable ranking views. Retry count remains available as a secondary table sort.

Rank configurations by

Score ranks high to low. Speed, cost and tokens rank low to high. Retry sorting remains available from the table header.

Results for all 21 model-mode configurations
ConfigurationSelected rankMean score / replicasLatencyCostTokens
Claude Fable 5Anthropic · official model page ↗#1by score100.0SD not available · single run113.3s$0.54811,7310.00
Claude Opus 5.5Anthropic · official model page ↗#1by score100.0SD not available · single run96.2s$0.23612,5900.00
DeepSeek V4 FlashDeepSeek · official model page ↗#1by score100.0SD not available · single run260.9s$0.008530,7200.00
DeepSeek V4 Flash (no thinking)DeepSeek · official model page ↗#1by score100.0SD not available · single run64.2s$0.00248,8290.00
DeepSeek V4 ProDeepSeek · official model page ↗#1by score100.0SD not available · single run167.8s$0.01112,8440.00
Gemini 3.5 FlashGoogle · official model page ↗#1by score100.0SD 0.0 · 3 retained runs474.6s$0.17019,4120.33
GPT-5.6 SolOpenAI · official model page ↗#1by score100.0SD not available · single run2750.9s$0.65422,3211.00
GPT-6 LunaOpenAI · official model page ↗#1by score100.0SD not available · single run144.3s$0.009118,7400.00
GPT-6.1 SolOpenAI · official model page ↗#1by score100.0SD not available · single run250.1s$0.11111,6070.00
Grok 4.5xAI · official model page ↗#1by score100.0SD not available · single run138.5s$0.0498,6980.00
Claude Sonnet 5.5Anthropic · official model page ↗#11by score99.5SD not available · single run67.4s$0.09910,6880.00
OpenAI o3OpenAI · official model page ↗#12by score97.9SD not available · single run61.7s$0.0739,5850.00
Gemma 4 31B ITGoogle · official model page ↗#13by score96.6SD not available · single run115.2s$0.00004,6970.00
Mistral Small 4Mistral · official model page ↗#13by score96.6SD not available · single run92.2s$0.01118,3690.00
DeepSeek V4.1 FlashDeepSeek · official model page ↗#15by score95.3SD not available · single run122.1s$0.03529,2630.00
Gemini 3.1 Pro PreviewGoogle · official model page ↗#16by score94.2SD not available · single run96.7s$0.14312,4360.00
Mistral Medium 3.5Mistral · official model page ↗#17by score92.6SD not available · single run109.6s$0.11515,8970.00
OpenAI GPT-4.1OpenAI · official model page ↗#18by score81.8SD not available · single run77.0s$0.0577,5550.00
Claude Sonnet 4.5Anthropic · official model page ↗#19by score77.9SD not available · single run55.4s$0.0544,2060.00
Grok 4.3xAI · official model page ↗#20by score69.5SD not available · single run96.1s$0.02811,4770.00
Claude Haiku 4.5Anthropic · official model page ↗#21by score64.5SD not available · single run31.1s$0.0245,2980.00
#1 by score

Claude Fable 5

Anthropic
100.0mean score

SD not available · single run

Latency
113.3s
Cost
$0.548
Tokens
11,731
Retries
0.00
Official model page ↗
#1 by score

Claude Opus 5.5

Anthropic
100.0mean score

SD not available · single run

Latency
96.2s
Cost
$0.236
Tokens
12,590
Retries
0.00
Official model page ↗
#1 by score

DeepSeek V4 Flash

DeepSeek
100.0mean score

SD not available · single run

Latency
260.9s
Cost
$0.0085
Tokens
30,720
Retries
0.00
Official model page ↗
#1 by score

DeepSeek V4 Flash (no thinking)

DeepSeek
100.0mean score

SD not available · single run

Latency
64.2s
Cost
$0.0024
Tokens
8,829
Retries
0.00
Official model page ↗
#1 by score

DeepSeek V4 Pro

DeepSeek
100.0mean score

SD not available · single run

Latency
167.8s
Cost
$0.011
Tokens
12,844
Retries
0.00
Official model page ↗
#1 by score

Gemini 3.5 Flash

Google
100.0mean score

SD 0.0 · 3 retained runs

Latency
474.6s
Cost
$0.170
Tokens
19,412
Retries
0.33
Official model page ↗
#1 by score

GPT-5.6 Sol

OpenAI
100.0mean score

SD not available · single run

Latency
2750.9s
Cost
$0.654
Tokens
22,321
Retries
1.00
Official model page ↗
#1 by score

GPT-6 Luna

OpenAI
100.0mean score

SD not available · single run

Latency
144.3s
Cost
$0.0091
Tokens
18,740
Retries
0.00
Official model page ↗
#1 by score

GPT-6.1 Sol

OpenAI
100.0mean score

SD not available · single run

Latency
250.1s
Cost
$0.111
Tokens
11,607
Retries
0.00
Official model page ↗
#1 by score

Grok 4.5

xAI
100.0mean score

SD not available · single run

Latency
138.5s
Cost
$0.049
Tokens
8,698
Retries
0.00
Official model page ↗
#11 by score

Claude Sonnet 5.5

Anthropic
99.5mean score

SD not available · single run

Latency
67.4s
Cost
$0.099
Tokens
10,688
Retries
0.00
Official model page ↗
#12 by score

OpenAI o3

OpenAI
97.9mean score

SD not available · single run

Latency
61.7s
Cost
$0.073
Tokens
9,585
Retries
0.00
Official model page ↗
#13 by score

Gemma 4 31B IT

Google
96.6mean score

SD not available · single run

Latency
115.2s
Cost
$0.0000
Tokens
4,697
Retries
0.00
Official model page ↗
#13 by score

Mistral Small 4

Mistral
96.6mean score

SD not available · single run

Latency
92.2s
Cost
$0.011
Tokens
18,369
Retries
0.00
Official model page ↗
#15 by score

DeepSeek V4.1 Flash

DeepSeek
95.3mean score

SD not available · single run

Latency
122.1s
Cost
$0.035
Tokens
29,263
Retries
0.00
Official model page ↗
#16 by score

Gemini 3.1 Pro Preview

Google
94.2mean score

SD not available · single run

Latency
96.7s
Cost
$0.143
Tokens
12,436
Retries
0.00
Official model page ↗
#17 by score

Mistral Medium 3.5

Mistral
92.6mean score

SD not available · single run

Latency
109.6s
Cost
$0.115
Tokens
15,897
Retries
0.00
Official model page ↗
#18 by score

OpenAI GPT-4.1

OpenAI
81.8mean score

SD not available · single run

Latency
77.0s
Cost
$0.057
Tokens
7,555
Retries
0.00
Official model page ↗
#19 by score

Claude Sonnet 4.5

Anthropic
77.9mean score

SD not available · single run

Latency
55.4s
Cost
$0.054
Tokens
4,206
Retries
0.00
Official model page ↗
#20 by score

Grok 4.3

xAI
69.5mean score

SD not available · single run

Latency
96.1s
Cost
$0.028
Tokens
11,477
Retries
0.00
Official model page ↗
#21 by score

Claude Haiku 4.5

Anthropic
64.5mean score

SD not available · single run

Latency
31.1s
Cost
$0.024
Tokens
5,298
Retries
0.00
Official model page ↗
04

Criterion-level benchmark record

Scoring evidence

Dimension and criterion values are descriptive results for the 21 evaluated configuration means. N/A remains distinct from zero.

Attainment scale0%50%100%N/A = not applicable
T193.7%Understanding / modelling
T294.8%Method selection
T392.4%Development
T490.5%Final results
T597.9%Domains / assumptions
T692.9%Consistency
T796.3%Rigor
T8N/AIndependent verification
T1 through T8 attainment for each model-mode configuration
ConfigurationT120 ptsT215 ptsT320 ptsT415 ptsT510 ptsT68 ptsT77 ptsT80 pts
Claude Fable 5100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
Claude Opus 5.5100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
DeepSeek V4 Flash100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
DeepSeek V4 Flash (no thinking)100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
DeepSeek V4 Pro100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
GPT-5.6 Sol100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
GPT-6 Luna100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
GPT-6.1 Sol100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
Gemini 3.5 Flash100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
Grok 4.5100.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
Claude Sonnet 5.5100.0%100.0%100.0%100.0%95.0%100.0%100.0%N/A
OpenAI o390.0%100.0%100.0%100.0%100.0%100.0%100.0%N/A
Gemma 4 31B IT90.0%100.0%100.0%91.7%100.0%100.0%100.0%N/A
Mistral Small 490.0%100.0%100.0%91.7%100.0%100.0%100.0%N/A
DeepSeek V4.1 Flash90.0%100.0%100.0%83.3%100.0%100.0%100.0%N/A
Gemini 3.1 Pro Preview85.0%100.0%100.0%83.3%100.0%100.0%100.0%N/A
Mistral Medium 3.595.0%100.0%87.5%91.7%100.0%81.3%89.3%N/A
OpenAI GPT-4.190.0%80.0%67.5%83.3%90.0%81.3%89.3%N/A
Claude Sonnet 4.585.0%76.7%72.5%63.3%90.0%87.5%78.6%N/A
Grok 4.375.0%75.0%60.0%50.0%90.0%62.5%89.3%N/A
Claude Haiku 4.577.5%58.3%52.5%61.7%90.0%37.5%75.0%N/A
05

Load on request

Retained responses and judgments

23 responses across 21 model-mode configurations. Long records load only when opened.

Interpretation boundary

Answers are reproduced as submitted. Each is paired with its authority-resolved judgment for this snapshot; this is not a general evaluation of the underlying model.

Compare two configurations All retained replicas

Select two configurations and load all their retained replicas.

Select two configurations and load all their retained replicas.

#1Claude Fable 5Anthropic · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1Claude Opus 5.5Anthropic · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1DeepSeek V4 FlashDeepSeek · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1DeepSeek V4 Flash (no thinking)DeepSeek · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1DeepSeek V4 ProDeepSeek · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1GPT-5.6 SolOpenAI · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1GPT-6 LunaOpenAI · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1GPT-6.1 SolOpenAI · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1Gemini 3.5 FlashGoogle · 3 retained runs · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#1Grok 4.5xAI · 1 retained run · load on request100.0configuration mean+
Configuration recordOfficial model page ↗
#11Claude Sonnet 5.5Anthropic · 1 retained run · load on request99.5configuration mean+
Configuration recordOfficial model page ↗
#12OpenAI o3OpenAI · 1 retained run · load on request97.9configuration mean+
Configuration recordOfficial model page ↗
#13Gemma 4 31B ITGoogle · 1 retained run · load on request96.6configuration mean+
Configuration recordOfficial model page ↗
#13Mistral Small 4Mistral · 1 retained run · load on request96.6configuration mean+
Configuration recordOfficial model page ↗
#15DeepSeek V4.1 FlashDeepSeek · 1 retained run · load on request95.3configuration mean+
Configuration recordOfficial model page ↗
#16Gemini 3.1 Pro PreviewGoogle · 1 retained run · load on request94.2configuration mean+
Configuration recordOfficial model page ↗
#17Mistral Medium 3.5Mistral · 1 retained run · load on request92.6configuration mean+
Configuration recordOfficial model page ↗
#18OpenAI GPT-4.1OpenAI · 1 retained run · load on request81.8configuration mean+
Configuration recordOfficial model page ↗
#19Claude Sonnet 4.5Anthropic · 1 retained run · load on request77.9configuration mean+
Configuration recordOfficial model page ↗
#20Grok 4.3xAI · 1 retained run · load on request69.5configuration mean+
Configuration recordOfficial model page ↗
#21Claude Haiku 4.5Anthropic · 1 retained run · load on request64.5configuration mean+
Configuration recordOfficial model page ↗
06

Record context

About this benchmark record

The Atlas presents the exercise, rubric, results and retained responses in one interactive record.

Exercise IDA3-25-P4-E1
SubjectAnalisi III
Exam sitting2025 · Fourth sitting
Benchmark snapshot2026-10-15
Rubric version1.0
Retained responses23
How to use this page

Use the Atlas to inspect what was tested and how each configuration performed. Most configuration–exercise cells are single-run, so displayed values are descriptive rather than uncertainty estimates.

UniAIBench v1.0Snapshot del benchmark live · 15 ottobre 2026Documento di lavoro · 14 agosto 2026Sito aggiornato · 15 ottobre 2026Codice sorgenteUniAIBench