AgenticBBO-Bench
Harbor: pendingA cross-domain benchmark evaluating LLM agents on five black-box optimization tasks (synthetic, HPO, DB tuning, chip design, molecular) under a unified finite-budget protocol.
- Samples
- N/A
- Released
- 2026
- Scores
- 7
RAP
RAP: A Dataset for Paper-to-Poster Generation via Declarative Markup from Human-Crafted Exemplars
A dataset and benchmark of 25K+ verified paper-poster pairs with HTML/CSS markup for objective render-and-compare evaluation of visual coding models.
How to run
ltb · rap
$ npm i -g lettokenburn-cli$ ltb bench info rap
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
5 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| Claude Opus 5 | 0.6 | N/A | Independent | medium | Oct 16, 2026 | arxiv.org |
| Qwen3.8 27B | 0.6 | N/A | Independent | medium | Oct 16, 2026 | arxiv.org |
| Gemini 3.1 Pro | 0.6 | N/A | Independent | medium | Oct 16, 2026 | arxiv.org |
| Kimi K3 | 0.6 | N/A | Independent | medium | Oct 16, 2026 | arxiv.org |
| GPT-6 Astra | 0.6 | N/A | Independent | medium | Oct 16, 2026 | arxiv.org |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsLoading community runs…
About
RAP contains 25,435 verified paper-poster pairs from ML conferences (2022-2026), annotated with content taxonomy and converted to editable HTML/CSS. It includes a held-out test set of 150 papers.
Evaluation uses an automated render-and-compare pipeline against human-crafted references, enabling objective verification rather than subjective VLM judging. The dataset supports retrieval-augmented generation and model training.