# Enterprise Task Execution on 10,000 Enterprise Tasks (test)

Source: [Enterprise Task Execution on 10,000 Enterprise Tasks (test)](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-12T07:30:19.698Z

Evaluation of model performance on executing tasks defined in the 10,000 Enterprise Tasks dataset (test split), focusing on functional success, safety adherence, operational efficiency (latency, cost), and violation rates.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Enterprise Task Execution |
| Dataset (including any stated split) | 10,000 Enterprise Tasks (test) |
| Metric | Functional Success Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Functional Success Rate](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Functional%20Success%20Rate>): Higher is better; unit: PERCENT.

- [Safe Success Rate](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Safe%20Success%20Rate>): Higher is better; unit: PERCENT.

- [Automated Safety Rate](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Automated%20Safety%20Rate>): Higher is better; unit: PERCENT.

- [Violations per 1k Tasks](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Violations%20per%201k%20Tasks>): Lower is better; unit: SCALAR.

- [Memory Poison Incidents per 1k Tasks](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Memory%20Poison%20Incidents%20per%201k%20Tasks>): Lower is better; unit: SCALAR.

- [Escalation Rate](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Escalation%20Rate>): Lower is better; unit: PERCENT.

- [Audit Coverage](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Audit%20Coverage>): Higher is better; unit: PERCENT.

- [Mean Handoffs](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Mean%20Handoffs>): Lower is better; unit: SCALAR.

- [Mean Cost](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=Mean%20Cost>): Lower is better; unit: SCALAR.

- [P95 Latency](<https://www.sota2.com/research/sota/enterprise-task-execution-on-10-000-enterprise-tasks-test.md?metric=P95%20Latency>): Lower is better; unit: TIME.

## Recorded results

| Method | Functional Success Rate | Result date | Report paper |
| --- | --- | --- | --- |
| CEAD proposed | 73 | Not recorded | 6a02b3f2dbc836b87d824558 |
| SOA-brokered agents | 63.1 | Not recorded | 6a02b3f2dbc836b87d824558 |
| Governance-first design-poor grid | 53.2 | Not recorded | 6a02b3f2dbc836b87d824558 |
| Prompt-first mono-agent | 52.4 | Not recorded | 6a02b3f2dbc836b87d824558 |
| Role-based micro-agent swarm | 30.8 | Not recorded | 6a02b3f2dbc836b87d824558 |

### CEAD proposed

Functional Success Rate: 73. Result ID: 6a02d6fd47f34ce924e545ef.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3f2dbc836b87d824558.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a02b3f2dbc836b87d824558.md>)

Recorded evaluation configuration: {"Architecture ID":"A4"}. This record does not assert independent reproduction.

### SOA-brokered agents

Functional Success Rate: 63.1. Result ID: 6a02d6fd47f34ce924e545ed.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3f2dbc836b87d824558.md>)

Recorded evaluation configuration: {"Architecture ID":"A2"}. This record does not assert independent reproduction.

### Governance-first design-poor grid

Functional Success Rate: 53.2. Result ID: 6a02d6fd47f34ce924e545ee.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3f2dbc836b87d824558.md>)

Recorded evaluation configuration: {"Architecture ID":"A3"}. This record does not assert independent reproduction.

### Prompt-first mono-agent

Functional Success Rate: 52.4. Result ID: 6a02d6fd47f34ce924e545eb.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3f2dbc836b87d824558.md>)

Recorded evaluation configuration: {"Architecture ID":"A0"}. This record does not assert independent reproduction.

### Role-based micro-agent swarm

Functional Success Rate: 30.8. Result ID: 6a02d6fd47f34ce924e545ec.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3f2dbc836b87d824558.md>)

Recorded evaluation configuration: {"Architecture ID":"A1"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.