Research Phase — FORGE is not production ready. Scores from [FLAGGED] categories are excluded from the main score. Some categories have known limitations.
OpenRouter, OpenAI, or any OpenAI-compatible model
Stored in browser only — never sent to any server except the API provider
5 questions · ~30s · <$0.01
Same seed = same questions for all models
10
Parallel API requests. Higher = faster but may hit rate limits.
Estimated seconds per API call. Used for runtime estimation.
Estimated runtime (quick mode): ~5 min
Connecting...
Category Breakdown
Category
Tiers
Score
Correct
Cliff Index
Research disclaimer: FORGE is in active research phase. Scores from [FLAGGED] categories (Polynomial Roots, RSA Arithmetic)
are excluded from the main score. Results from [RESEARCH] categories should be interpreted with caution.
Current results are preliminary.
Multi-Model Comparison
Run multiple models with the same seed, then compare their results side by side.
Starting...
Score Comparison
Interpolation vs Extrapolation
Category × Model Heatmap
Model ↕
FORGE Score ↕
Interpolation ↕
Extrapolation ↕
Cliff Index ↕
Cost ↕
Time ↕
Saved Snapshots
Snapshots are saved in your browser's localStorage. Export to file for backup or sharing.
∅
No snapshots saved yet. Run a benchmark and save the result.