---
title: "Is the expensive model worth it? | OpenRA arena investigation | Evals by Jetty"
url: https://evaljetty.com/openra/investigations/cost-vs-outcome.html
description: "Not in this field: Jev Router delivered the cheapest wins ($0.17 per win), and per-game cost ranged from $0.03 to $0.47."
updated: 2026-09-28
publisher: Jetty (https://jetty.io)
---

03 — Investigation · Is it worth the cost?

# Is the expensive model worth it?

24 runsRound 230-minute wall-clock budget per gamerunbook v2.0.0 (evals v2)collection jettygrowthteamtask openra-arena-matchupdated 2026-09-28

**Decision it informs:** Which model to pay for when running many games (or many agent tasks like them).

## Key takeaway generated

Not in this field: Jev Router delivered the cheapest wins ($0.17 per win), and per-game cost ranged from $0.03 to $0.47.

## At a glance computed

Runs

24

one game per Jetty run

Model spend

$11.69

OpenRouter, both players

Sandbox time

4.2 h

wall clock, all runs

Eval pass rate

96%

23/24 runs passed

## Key figure computed

### Cost per game vs win rate

Each dot is one model over its round-robin games. Up and to the left is better.

Table view

| Model | Cost / game | Win rate | Cost / win |
| --- | --- | --- | --- |
| Jev Router | $0.11 | 67% | $0.17 |
| Claude Sonnet 5 | $0.36 | 67% | $0.54 |
| GPT-5.6 Terra | $0.47 | 67% | $0.70 |
| GPT-5.6 Luna | $0.03 | 0% | – |

## Evaluations computed

Every run's evals-v2 validation report, scored by Jetty: deterministic code checks and the reviewer checklist.

| Evaluation | Passed | Rate | Share |
| --- | --- | --- | --- |
| Code checks | 167/168 | 99% |  |
| Checklist items | 72/72 | 100% |  |
| Runs with a passing verdict | 23/24 | 96% |  |

## Findings generated

### Jev Router buys a win for the least money

medium confidence

$0.17 of model spend per game won (8 wins at $0.11 a game). Jev Router $0.17; Claude Sonnet 5 $0.54; GPT-5.6 Terra $0.70; GPT-5.6 Luna no wins.

24 round-robin gamescost\_usd per run

### Cost per game spans 14×

high confidence

From $0.03 (GPT-5.6 Luna) to $0.47 (GPT-5.6 Terra) per game at the same number of turns.

Jev Router $0.11Claude Sonnet 5 $0.36GPT-5.6 Terra $0.47GPT-5.6 Luna $0.03

### The winner is also not the most expensive

medium confidence

Jev Router topped the table at $0.11 a game. Spend and strength are only loosely related in this field.

Jev Router 8–4

## Proposed improvements generated

Runbook

### Trim the briefing

Most tokens are the per-turn briefing; summarizing idle units and static map facts would cut prompt cost for every model.

backed by: briefings in decisions.jsonl

Run

### Add the cheapest strong model to the ladder

Run Jev Router against the built-in AI too, to see if the best value also holds up there.

backed by: $0.17 per win

## Caveats

- Costs are OpenRouter charges for the competing models only, not the runbook agent or the sandbox.
- Router pricing varies per request.
- 4 games per pair.

**Runs and provenance**

| Run | Stage | West | East | Winner | Evaluation | Game min | Wall min | Model cost | Jetty run ID |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| [r2-rr-jev-gpt56terra-g1](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56terra-g1.html) | Round robin | Jev Router | GPT-5.6 Terra | GPT-5.6 Terra | failchecks **9/10** | 10.2 | 17.0 | $0.69 | 76e96f3a |
| [r2-rr-gpt56terra-gpt56luna-g4](https://evaljetty.com/openra/runs/r2-rr-gpt56terra-gpt56luna-g4.html) | Round robin | GPT-5.6 Luna | GPT-5.6 Terra | GPT-5.6 Terra | passchecks **10/10** | 6.1 | 7.8 | $0.36 | 07fd5299 |
| [r2-rr-gpt56terra-gpt56luna-g3](https://evaljetty.com/openra/runs/r2-rr-gpt56terra-gpt56luna-g3.html) | Round robin | GPT-5.6 Terra | GPT-5.6 Luna | GPT-5.6 Terra | passchecks **10/10** | 7.8 | 10.2 | $0.50 | 77076920 |
| [r2-rr-gpt56terra-gpt56luna-g2](https://evaljetty.com/openra/runs/r2-rr-gpt56terra-gpt56luna-g2.html) | Round robin | GPT-5.6 Luna | GPT-5.6 Terra | GPT-5.6 Terra | passchecks **10/10** | 5.9 | 7.7 | $0.36 | 7caea988 |
| [r2-rr-sonnet5-gpt56luna-g1](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56luna-g1.html) | Round robin | Claude Sonnet 5 | GPT-5.6 Luna | Claude Sonnet 5 | passchecks **10/10** | 6.1 | 8.6 | $0.34 | 7fa30839 |
| [r2-rr-sonnet5-gpt56terra-g4](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56terra-g4.html) | Round robin | GPT-5.6 Terra | Claude Sonnet 5 | Claude Sonnet 5 | passchecks **10/10** | 8.7 | 12.4 | $0.91 | fc8c2b35 |
| [r2-rr-gpt56terra-gpt56luna-g1](https://evaljetty.com/openra/runs/r2-rr-gpt56terra-gpt56luna-g1.html) | Round robin | GPT-5.6 Terra | GPT-5.6 Luna | GPT-5.6 Terra | passchecks **10/10** | 4.7 | 6.0 | $0.28 | 53107caf |
| [r2-rr-sonnet5-gpt56luna-g4](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56luna-g4.html) | Round robin | GPT-5.6 Luna | Claude Sonnet 5 | Claude Sonnet 5 | passchecks **10/10** | 5.4 | 6.7 | $0.30 | a816405d |
| [r2-rr-sonnet5-gpt56luna-g3](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56luna-g3.html) | Round robin | Claude Sonnet 5 | GPT-5.6 Luna | Claude Sonnet 5 | passchecks **10/10** | 7.2 | 10.1 | $0.43 | 67220f81 |
| [r2-rr-sonnet5-gpt56luna-g2](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56luna-g2.html) | Round robin | GPT-5.6 Luna | Claude Sonnet 5 | Claude Sonnet 5 | passchecks **10/10** | 5.9 | 7.8 | $0.32 | a5fa0d15 |
| [r2-rr-sonnet5-gpt56terra-g3](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56terra-g3.html) | Round robin | Claude Sonnet 5 | GPT-5.6 Terra | GPT-5.6 Terra | passchecks **10/10** | 4.8 | 7.3 | $0.50 | c243bf0a |
| [r2-rr-sonnet5-gpt56terra-g2](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56terra-g2.html) | Round robin | GPT-5.6 Terra | Claude Sonnet 5 | Claude Sonnet 5 | passchecks **10/10** | 11.7 | 16.8 | $1.32 | bcce8dad |
| [r2-rr-sonnet5-gpt56terra-g1](https://evaljetty.com/openra/runs/r2-rr-sonnet5-gpt56terra-g1.html) | Round robin | Claude Sonnet 5 | GPT-5.6 Terra | GPT-5.6 Terra | passchecks **10/10** | 6.1 | 8.4 | $0.61 | 410f87f9 |
| [r2-rr-jev-gpt56luna-g4](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56luna-g4.html) | Round robin | GPT-5.6 Luna | Jev Router | Jev Router | passchecks **10/10** | 5.1 | 7.9 | $0.11 | a5d0f3fc |
| [r2-rr-jev-gpt56luna-g3](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56luna-g3.html) | Round robin | Jev Router | GPT-5.6 Luna | Jev Router | passchecks **10/10** | 5.4 | 7.9 | $0.10 | 9e9a7d94 |
| [r2-rr-jev-gpt56luna-g2](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56luna-g2.html) | Round robin | GPT-5.6 Luna | Jev Router | Jev Router | passchecks **10/10** | 4.9 | 6.6 | $0.08 | c5d4707d |
| [r2-rr-jev-gpt56luna-g1](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56luna-g1.html) | Round robin | Jev Router | GPT-5.6 Luna | Jev Router | passchecks **10/10** | 5.6 | 8.3 | $0.12 | 3531a1f4 |
| [r2-rr-jev-gpt56terra-g4](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56terra-g4.html) | Round robin | GPT-5.6 Terra | Jev Router | Jev Router | passchecks **10/10** | 9.3 | 14.9 | $0.62 | 6d4b5d06 |
| [r2-rr-jev-gpt56terra-g3](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56terra-g3.html) | Round robin | Jev Router | GPT-5.6 Terra | Jev Router | passchecks **10/10** | 14.6 | 21.8 | $1.11 | eda9e5a6 |
| [r2-rr-jev-gpt56terra-g2](https://evaljetty.com/openra/runs/r2-rr-jev-gpt56terra-g2.html) | Round robin | GPT-5.6 Terra | Jev Router | GPT-5.6 Terra | passchecks **10/10** | 9.5 | 15.2 | $0.74 | 9872e42e |
| [r2-rr-jev-sonnet5-g4](https://evaljetty.com/openra/runs/r2-rr-jev-sonnet5-g4.html) | Round robin | Claude Sonnet 5 | Jev Router | Claude Sonnet 5 | passchecks **10/10** | 5.6 | 8.5 | $0.35 | 55fe7dbc |
| [r2-rr-jev-sonnet5-g3](https://evaljetty.com/openra/runs/r2-rr-jev-sonnet5-g3.html) | Round robin | Jev Router | Claude Sonnet 5 | Jev Router | passchecks **10/10** | 11.5 | 17.9 | $0.77 | 188481ef |
| [r2-rr-jev-sonnet5-g2](https://evaljetty.com/openra/runs/r2-rr-jev-sonnet5-g2.html) | Round robin | Claude Sonnet 5 | Jev Router | Jev Router | passchecks **10/10** | 5.0 | 7.2 | $0.29 | dcf50a3a |
| [r2-rr-jev-sonnet5-g1](https://evaljetty.com/openra/runs/r2-rr-jev-sonnet5-g1.html) | Round robin | Jev Router | Claude Sonnet 5 | Claude Sonnet 5 | passchecks **10/10** | 6.9 | 10.1 | $0.48 | 80bb2b04 |

---
Source: https://evaljetty.com/openra/investigations/cost-vs-outcome.html · Evals by Jetty · Run your own evals: https://jetty.io/?utm_source=evaljetty&utm_medium=llms
