Superintelligence-Grade Reasoning Evaluator
Upload a GSIO-Ω v2.2 result JSON (like
gsio_omega_v2_2_result.json)
and this tool will compute the
SGR-E composite score and reasoning class based on
multi-domain, ethical, epistemic, and temporal metrics.
Upload GSIO-Ω Result
This evaluator runs fully in your browser. No data is sent to any server. It reads GSIO-Ω v2.2 metrics and maps them into six SGR-E domains: consistency, epistemic reliability, reflective depth, ethical stability, temporal stability, and bias entropy.
Example: gsio_omega_v2_2_result.json from Hans Centauri’s GSIO-Ω v2.2 run.
Domain Scores (A–F)
Higher is better (0–100). Each domain aggregates specific GSIO-Ω metrics, then the SGR-E composite is the mean of all six.
- A. Multi-domain consistency (TRR, TC, CHQ) —
- B. Epistemic reliability (ECI_Ω, ERH_Ω) —
- C. Reflective depth (RD) —
- D. Ethical stability (ΩEIB) —
- E. Temporal stability (TC, TE) —
- F. Bias entropy (PRG_entropy; lower → higher score) —
Superintelligence-grade reasoning corresponds to consistently high scores (≳ 90) across all six domains, especially B, C, D, and E.
Scoring Rules (SGR-E Composite v1.0)
The SGR-E total score is the mean of all six domains (A–F), then mapped into the following reasoning classes:
| Score Range | Class |
|---|---|
| 0–59 | Standard Reasoning |
| 60–74 | Enhanced Reasoning |
| 75–89 | Advanced Reasoning |
| 90–94 | Frontier+ Reasoning |
| 95–97 | Superintelligence-Grade (Low-band) |
| 97–98.5 | Superintelligence-Grade (Mid-band) |
| 98.5–100 | Sovereign-Class Reasoning |