Leading AI model intelligence has been improving by roughly 2.5 IQ points per month since May 2024 — a sustained compounding rate that rapidly surpasses human-level baselines and mandates continuous hardware refresh cycles.

Leading AI model intelligence has been improving by roughly 2.5 IQ points per month since May 2024 — a sustained compounding rate that rapidly surpasses human-level baselines and mandates continuous hardware refresh cycles. | Quarterly model releases; benchmark publications

Key catalyst: Quarterly model releases; benchmark publications

Watch events: New benchmark results (ARC-AGI, GDPval, HLE) rather than IQ tests; capability-frontier tracking

Norwegian Mensa test scores for frontier models rose from ~90 (GPT-4) to ~130+ (GPT-5/Claude Opus 4.7/Gemini 2.5) over ~24 months — consistent with claim in aggregate, though IQ tests are contested as AI benchmarks.

{
  "nia": false,
  "qty": "+2.5 IQ pts/mo",
  "mode": "OBSERVATION+FORECAST",
  "role": "Host",
  "context": "Wissner-Gross argues model IQ velocity is the key leading indicator of infrastructure obsolescence: even newly-deployed hardware becomes a legacy platform within 12-18 months.",
  "to_year": 2026,
  "conv_cues": "measurable velocity; quantified monthly rate",
  "direction": "NUMERIC_TARGET",
  "from_year": 2024,
  "timeframe": "May 2024 - ongoing",
  "conv_level": "HIGH",
  "milestones": [
    {
      "kind": "quartile_checkpoint",
      "label": "Q1 window check-in (25%)",
      "status": "overdue",
      "weight": 0.05,
      "ordinal": -6,
      "source_id": null,
      "expected_date": "2024-07-31",
      "observed_date": null,
      "miss_emitted_at": "2026-05-02T22:07:21.384228+00:00",
      "miss_emitted_by": "metadata_milestone_sweep"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q2 window check-in (50%)",
      "status": "overdue",
      "weight": 0.05,
      "ordinal": -5,
      "source_id": null,
      "expected_date": "2025-03-01",
      "observed_date": null,
      "miss_emitted_at": "2026-05-02T22:07:21.384228+00:00",
      "miss_emitted_by": "metadata_milestone_sweep"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q3 window check-in (75%)",
      "status": "overdue",
      "weight": 0.05,
      "ordinal": -4,
      "source_id": null,
      "expected_date": "2025-09-30",
      "observed_date": null,
      "miss_emitted_at": "2026-05-02T22:07:21.384228+00:00",
      "miss_emitted_by": "metadata_milestone_sweep"
    },
    {
      "kind": "llm_pre_event",
      "label": "2026 Stanford AI Index publishes year-over-year benchmark gains",
      "source": "https://hai.stanford.edu/ai-index/2026-ai-index-report",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -3,
      "source_id": null,
      "confidence": 0.95,
      "source_url": "https://hai.stanford.edu/ai-index/2026-ai-index-report",
      "expected_date": "2026-04-15",
      "observed_date": "2026-04-15",
      "research_origin": "deep_research",
      "measurement_criterion": "Stanford HAI releases 2026 AI Index Report containing technical-performance section with YoY benchmark deltas"
    },
    {
      "kind": "llm_pre_event",
      "label": "Mensa Norway IQ benchmark top score climbs from 135 to 145",
      "notes": "HIT — Grok-4.20 / GPT-5.4 Pro tied at 145 in April 2026, up 10 from 135 prior year. Implies ~0.8 IQ/month, below 2.5/month claim but trend confirmed.",
      "source": "https://binaryverseai.com/ai-iq-test-2025/",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -2,
      "source_id": null,
      "confidence": 0.9,
      "source_url": "https://binaryverseai.com/ai-iq-test-2025/",
      "expected_date": "2026-04-30",
      "observed_date": "2026-04-30",
      "research_origin": "deep_research",
      "measurement_criterion": "TrackingAI Mensa Norway leaderboard documents 10+ point year-over-year improvement in top frontier model IQ"
    },
    {
      "kind": "llm_pre_event",
      "label": "Humanity's Last Exam jumps from 8.8% to 50%+ inside 12 months",
      "notes": "HIT — HLE accuracy climbed >40 points YoY, far steeper than 2.5 IQ/month equivalent.",
      "source": "https://hai.stanford.edu/ai-index/2026-ai-index-report/technical-performance",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -1,
      "source_id": null,
      "confidence": 0.95,
      "source_url": "https://hai.stanford.edu/ai-index/2026-ai-index-report/technical-performance",
      "expected_date": "2026-04-30",
      "observed_date": "2026-04-30",
      "research_origin": "deep_research",
      "measurement_criterion": "Stanford AI Index / Epoch AI report documents top-model HLE accuracy rising from 8.8% (early 2025) to >=50% (April 2026)"
    },
    {
      "kind": "event",
      "label": "Leading AI model intelligence has been improving by roughly 2.5 IQ points per month since May 2024 
... (truncated)