Recent Papers / arXiv:2605.13950
Collider-Bench: Benchmarking AI Agents with Particle Physics Analysis Reproduction
Darius A. Faroughy, Sofia Palacios Schweitzer, Ian Pang, Siddharth Mishra-Sharma, David Shih
Abstract
Collider-Bench is a benchmark for evaluating autonomous LLM agents on long-horizon, real-world scientific tasks involving the reproduction of Large Hadron Collider (LHC) experimental analyses.
Agents must turn published papers into executable simulation-and-selection pipelines to predict collision event yields, evaluated against quantitative targets.
Tasks
editResults
6 results reproduced from this paper.
Sorted instantly in-page
Results
6
SOTA rows
1
Models
6
Datasets
1
| # | Model | Vendor | Benchmark | Value | SOTA | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | Codex CLI (GPT-5.5) | OpenAI | Collider-Bench | 30.00 | #1 | — | source ↗ |
| 02 | Claude Code (Opus 4.7) | Anthropic | Collider-Bench | 20.00 | — | source ↗ | |
| 03 | Claude Code (Sonnet 4.6) | Anthropic | Collider-Bench | 10.00 | — | source ↗ | |
| 04 | Claude Code (Haiku 4.5) | Anthropic | Collider-Bench | 0.00 | — | source ↗ | |
| 05 | Codex CLI (GPT-5.4-mini) | OpenAI | Collider-Bench | 0.00 | — | source ↗ | |
| 06 | ForgeCode (DeepSeek-V4) | DeepSeek | Collider-Bench | 0.00 | — | source ↗ |
CodeSOTA extraction
Benchmark evidence
Link this paper to benchmark rows, datasets, model cards, and reproduced results as evidence is extracted.