AgentHorizon
Harbor: pendingA benchmark of 1,373 computer-use tasks with paired instructions to evaluate agentic judges on long-horizon trajectory success verification across three OSs.
- Samples
- 1,373
- Released
- 2026
- Scores
- 10
CroissantMiner
CroissantMiner Benchmark
CroissantMiner A benchmark for extracting Croissant metadata from ML dataset papers, released with the NeurIPS 2026 paper CroissantMiner: Automated Extraction and Validation of Croissant Metadata for ML Datasets (Evaluations and Datasets track).
How to run
ltb · croissantminer
$ npm i -g lettokenburn-cli$ ltb bench info croissantminer
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
11 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| Claude Sonnet 4.6 | 0.7 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| Claude Opus 4.7 | 0.7 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| GPT-5.4 | 0.7 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| Qwen3.6 35B A3B | 0.6 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| GLM-5.1 | 0.6 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| Gemini 2.5 Flash | 0.6 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| GPT-5.4-mini | 0.6 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| Gemini 3.1 Pro | 0.6 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| DeepSeek V3.2 | 0.6 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| Mistral Small 4 | 0.5 | single-pass | Independent | medium | Oct 5, 2026 | arxiv.org |
| Llama 4 Scout 17B | 0.4 | single-pass | Independent | medium | Oct 5, 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
CroissantMiner
A benchmark for extracting Croissant metadata from ML dataset papers, released with the NeurIPS 2026 paper CroissantMiner: Automated Extraction and Validation of Croissant Metadata for ML Datasets (Evaluations and Datasets track). Code: github.com/berkearda/croissantminer · Leaderboard: berkearda.github.io/croissantminer · Demo: Hugging Face Space · Paper: arXiv:2610.07132 · Version reviewed at NeurIPS: croissantminer/croissantminer (tag v1.0)
Contents…