Every run behind the Holo4 benchmark scores: 7,366 agent trajectories from Holo4 27B and Holo4 35B-A3B, with each step's reasoning, actions, tool results and screenshots.
Browse them at trajectories.hcompany.ai. This dataset is the same bundle the site serves.
| Benchmark | Holo4 27B | Holo4 35B-A3B | Upstream | License |
|---|---|---|---|---|
| OSWorld | 1,096 | 1,102 | xlang-ai/OSWorld | Apache 2.0 |
| OSWorld 2 | 106 | 106 | xlang-ai/OSWorld-V2 | Apache 2.0 |
| AndroidWorld | 344 | 347 | google-research/android_world | Apache 2.0 |
| AutomationBench | 1,600 | 1,598 | zapier/AutomationBench | MIT |
| PinchBench | 429 | 429 | pinchbench/skill | MIT |
| Agents' Last Exam | 104 | 105 | rdi-berkeley/agents-last-exam | Apache 2.0 |
data/index.json benchmarks and one summary row per trajectory
data/t/<id>.json full trajectory: task, steps, verifier result, token usage
img/<id>/<step>.webp screenshots, referenced from the steps
Each row in data/index.json has id, model, benchmark, run, task, instruction, success, score, duration_s and steps.
<PII removed>. Screenshots that showed them are replaced by a placeholder, and a few tasks are left out.Apache 2.0. Upstream task content keeps its original license.
Every run behind the Holo4 benchmark scores: 7,366 agent trajectories from Holo4 27B and Holo4 35B-A3B, with each step's reasoning, actions, tool results and screenshots.
Browse them at trajectories.hcompany.ai. This dataset is the same bundle the site serves.
| Benchmark | Holo4 27B | Holo4 35B-A3B | Upstream | License |
|---|---|---|---|---|
| OSWorld | 1,096 | 1,102 | xlang-ai/OSWorld | Apache 2.0 |
| OSWorld 2 | 106 | 106 | xlang-ai/OSWorld-V2 | Apache 2.0 |
| AndroidWorld | 344 | 347 | google-research/android_world | Apache 2.0 |
| AutomationBench | 1,600 | 1,598 | zapier/AutomationBench | MIT |
| PinchBench | 429 | 429 | pinchbench/skill | MIT |
| Agents' Last Exam | 104 | 105 | rdi-berkeley/agents-last-exam | Apache 2.0 |
data/index.json benchmarks and one summary row per trajectory
data/t/<id>.json full trajectory: task, steps, verifier result, token usage
img/<id>/<step>.webp screenshots, referenced from the steps
Each row in data/index.json has id, model, benchmark, run, task, instruction, success, score, duration_s and steps.
<PII removed>. Screenshots that showed them are replaced by a placeholder, and a few tasks are left out.Apache 2.0. Upstream task content keeps its original license.