# Cohere study compares coding agent team strategies

> In Cohere’s ProgramBench tests, planning-focused teams reached similar scores at lower reported cost.

- Canonical URL: https://newsverum.com/en/news/cohere-study-finds-agent-team-planning-efficiency
- Publisher: Newsverum
- Section: Technology
- Country: Canada
- Published: 2026-10-15T16:53:08+00:00
- Updated: 2026-10-15T16:53:08+00:00
- Event / data period: 2026-10-14
- Other editions: [Türkçe](https://newsverum.com/tr/news/cohere-study-finds-agent-team-planning-efficiency.md), [العربية](https://newsverum.com/ar/news/cohere-study-finds-agent-team-planning-efficiency.md), [Français](https://newsverum.com/fr/news/cohere-study-finds-agent-team-planning-efficiency.md), [Español](https://newsverum.com/es/news/cohere-study-finds-agent-team-planning-efficiency.md), [Deutsch](https://newsverum.com/de/news/cohere-study-finds-agent-team-planning-efficiency.md), [Português](https://newsverum.com/pt/news/cohere-study-finds-agent-team-planning-efficiency.md)

Cohere researchers published a controlled study comparing ways to organize coding-agent teams. On 200 ProgramBench tasks, an asynchronous GPT-5.6 Sol team delegating only planning reported 63.6% reward at a cost of $249; a GPT-5.6 Sol team delegating all stages scored 63.2% at $510. Both figures were averages across three runs.

A GPT-5.6 Sol lead coordinating 16 GPT-5.6 Luna planning workers scored 61.8% at $155, compared with 63.6% at $249 for the all-Sol planning team. This fixed 16-worker configuration was run once. Cohere says the findings are specific to its test harness; some model comparisons used a single trial, and the results do not establish general agent performance.

## Data

### Cohere study compares coding agent team strategies

Date and scope of the development reported in the primary source.

Methodology: The announcement date published by the source is recorded as one event.

| date | label | description | status |
|---|---|---|---|
| 2026-10-14 | Cohere study compares coding agent team strategies | In Cohere’s ProgramBench tests, planning-focused teams reached similar scores at lower reported cost. | occurred |

Sources: <https://cohere.com/blog/lessons-from-scaling-teams-of-frontier-coding-agents> · Download data: [JSON](https://newsverum.com/api/v1/articles/cohere-study-finds-agent-team-planning-efficiency/data/announcement-event?lang=en&format=json) · [CSV](https://newsverum.com/api/v1/articles/cohere-study-finds-agent-team-planning-efficiency/data/announcement-event?lang=en&format=csv)

### Cohere ProgramBench team results

ProgramBench results from Cohere’s coding-agent team evaluation; reward and inference cost are reported as stated by the source.

Methodology: Reward is the mean share of hidden behavioral tests passed across 200 ProgramBench tasks. The first two asynchronous GPT-5.6 Sol configurations average three runs; the fixed configuration with 16 GPT-5.6 Luna workers was run once. Results are from Cohere’s own test harness.

| Team configuration | Reward score | Total inference cost (USD) | Trial design |
|---|---|---|---|
| GPT-5.6 Sol; async lead delegates planning only | 63.6 | 249 | 200 tasks; mean of 3 runs |
| GPT-5.6 Sol; async lead delegates all stages | 63.2 | 510 | 200 tasks; mean of 3 runs |
| GPT-5.6 Sol lead; 16 GPT-5.6 Luna planning workers | 61.8 | 155 | Fixed 16-worker configuration; 1 run |

Sources: <https://cohere.com/blog/lessons-from-scaling-teams-of-frontier-coding-agents> · Download data: [JSON](https://newsverum.com/api/v1/articles/cohere-study-finds-agent-team-planning-efficiency/data/programbench-team-results?lang=en&format=json) · [CSV](https://newsverum.com/api/v1/articles/cohere-study-finds-agent-team-planning-efficiency/data/programbench-team-results?lang=en&format=csv)

## Sources

- [Cohere](<https://cohere.com/blog/lessons-from-scaling-teams-of-frontier-coding-agents>) (Source checked: 2026-10-15; Published: 2026-10-14)

---

Cite the canonical URL above. [Editorial standards and corrections](https://newsverum.com/en/editorial)