---
license: cc-by-4.0
language:
- en
task_categories:
- other
tags:
- scientific-agents
- benchmark
- evaluation
pretty_name: FrontierChallenge
configs:
- config_name: default
  data_files:
  - split: train
    path: manifest.jsonl
---

# FrontierChallenge

FrontierChallenge provides **97 scientific workflow tasks** with plaintext
English instructions, inputs, Harbor definitions, domain labels, and the
redistributable open runtime image.

| Path | Contents |
|---|---|
| `manifest.jsonl` | Dataset Viewer rows with task ID, taxonomy, difficulty, runtime, and instruction |
| `tasks/<task-id>/` | `instruction.md`, task metadata, environment definition, and agent-visible inputs |
| `images/` | Verified `linux/amd64` Docker archive for the 81 open-image tasks |

This repository contains no graders, rubrics, fixtures, or reference outputs.
Those are stored as encrypted archives in the separate
[`apodex/FrontierChallenge-reference`](https://huggingface.co/datasets/apodex/FrontierChallenge-reference)
dataset.

Use the [FrontierChallenge runtime](https://github.com/ApodexAI/FrontierAgent/tree/main/benchmarks/frontierchallenge)
to download both datasets, verify their shared registry, load the image archive,
run Harbor, and score a task. The evaluator host needs Python 3.12+ for the
pinned Harbor 0.20.0; task containers keep their own frozen Python versions.

```bash
git clone https://github.com/ApodexAI/FrontierAgent.git
cd FrontierAgent/benchmarks/frontierchallenge
python3.12 -m venv .venv
source .venv/bin/activate
python -m pip install -e .
HF_TOKEN=hf_... ./scripts/setup.sh --track open
cp .env.example .env
# Fill model and judge credentials in .env before running.
./scripts/run_eval.sh \
  --agent claude-code --model <model> \
  --include-task-name task_011_cell_migration_wound_healing
```

The 16 task definitions that execute ORCA and their inputs are included
normally. ORCA itself and an ORCA-configured image are never distributed;
evaluators obtain ORCA from its official provider and follow the runtime
repository's [local-image tutorial](https://github.com/ApodexAI/FrontierAgent/blob/main/benchmarks/frontierchallenge/docs/providers/orca.md).

Track membership follows what a task executes, not software named in supplied
files. For example, `task_098_orca_claisen_thermochemistry` reads precomputed
ORCA output without running ORCA, so it remains in the open track.

Verify a downloaded solve package with:

```bash
python tools/verify_dataset.py
```

## Metrics

Official Pass Rate uses a **strict score threshold**: a task passes only when
`evaluation_complete == 1` and valid `task_score > 0.999`. Native per-task
thresholds do not determine this metric.
Use unrounded scores: exactly `0.999` does not pass; `0.9991` and `1.0` pass.

- **Pass Rate (%)** = 100 × number of completed tasks with valid `task_score > 0.999` / 97.
- **Score (0–100)** = 100 × sum of valid, completed `task_score` values / 97.
- Partial credit remains the native rubric score normalized to [0, 1].
  Missing tasks, invalid scores, and incomplete evaluations contribute zero
  to the fixed denominator and must be reported separately; incomplete runs
  are not final benchmark results.
- Use one predeclared attempt per task. A subset may use its predeclared task
  count as denominator, but must be labeled as a subset, not the 97-task result.

There is only one pass field, `passed`, using the same strict threshold in
`reward.json`, `summary.csv`, and `summary.json`. No alternate pass field is
emitted. The runtime applies this rule to the staged reward adapter after
unsealing; encrypted archives and partial-credit rubrics remain unchanged.
The updated runtime also removes native pass decisions from published verifier
diagnostics. Jobs created by older runtimes cannot be resumed under the new
policy: use a fresh job name. Old files are left untouched; summarizing an old
job recomputes the official metric but does not rewrite its logs or rewards.
The summary records `metric_definition: score-gt-0.999`. Historical results
using per-task thresholds, `== 1.0`, or `>= 0.999` must be recomputed from raw
scores before comparison. See the
[scoring guide](https://github.com/ApodexAI/FrontierAgent/blob/main/benchmarks/frontierchallenge/docs/scoring.md)
and [runtime update](https://github.com/ApodexAI/FrontierAgent/pull/30).

**Runtime rollout:** the runtime update is pending. Before reporting results,
check that your summary records `metric_definition: score-gt-0.999`;
older runtimes use the legacy rule. Use the runtime summary for the fixed
97-task denominator; Harbor may aggregate only attempted trials.

## Integrity

The root `README.md` is a mutable dataset card and is intentionally outside
`checksums.sha256`. Task files (including task-level READMEs), manifests,
registries, and verification tools remain checksummed. Payload changes require
regenerating their checksum entries; editing this card does not.

## Citation

```bibtex
@misc{apodex11,
  title         = {Apodex 1.1: Scaling Agentic Intelligence for Complex Work},
  author        = {{Apodex Team}},
  year          = {2026},
  eprint        = {2608.23283},
  archivePrefix = {arXiv},
  primaryClass  = {cs.AI},
  url           = {https://arxiv.org/abs/2608.23283}
}

@misc{frontierchallenge,
  title         = {FrontierChallenge: Evaluating Scientific Workflow Completion},
  author        = {Liangcai Su and Zhaopeng Feng and Zhuo Chen and Zhen Zhang
                   and Xiang Lin and Ruilin Li and Handuo Zhang and Ning Wang
                   and Kailong Wen and Yueqi Guo and Feng Xing and Yiling Guo
                   and Chenxiong Qian and Simon Shaolei Du and Lidong Bing
                   and Xinyu Wang},
  year          = {2026},
  eprint        = {2608.24979},
  archivePrefix = {arXiv},
  primaryClass  = {cs.AI},
  url           = {https://arxiv.org/abs/2608.24979}
}
```