Interactive Program Induction - 100 Problems Selection - Minimax, unif.: leaderboard
Pinned appendix A5.T8; same released instruction-tuned models; exact subset/query/prior protocol in caption. A4T3 three seeds, evaluator rebuiltfromsinglevisibleexample. A5T7 100problems, temp1, three weightingrules. A5T8 100problems, nativeprompttemperatures andfixedselectionrule. A5T9 112sampledprograms, nooraclequeries. No cost/query-count metrics or duplicate Ours columns. Metric: Program accuracy (0–1). Source: https://arxiv.org/html/2611.18485v1
Metric: Program accuracy (0–1). Source: arxiv.org. Saturation forecast: Estimated already saturated. 3 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 32B Instruct | 0.293 | #1257 |
| 2 | Llama 3.1 8B Instruct | 0.195 | #4576 |
| 3 | Llama 3.2 3B Instruct | 0.146 | #7230 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=interactive-program-induction-100-problems-selection-minimax-unif · How It Works · Data refreshed daily, snapshot 2026-11-17.