Model · Google
Google Gemini Flash 3 (preview)
6 appearances · meta-score 61.8/100 · last seen 2026-07-10 · scored under rubric v15
Pooled composite across 6 primary runs: 48.2 ± 12.1 (95% CI)
- $197.27Total cost
- 639,225,183Tokens
- 48.8%Cache hit
- 10,778Requests
- 0.32%Failure rate
- 1.05Tools/response
- 3,240,372Tokens per $
Agent role scores
| Agent role | Score / 10 |
|---|---|
| Scout Agent | 7.63 |
| Sprint-Plan Agent | 7.48 |
| Spec Agent | 7.15 |
| Architect | 7.12 |
| Test Agent | 6.87 |
| Pseudocode Agent | 6.17 |
| Sprint-Review Agent | 6.13 |
| Platform Agent | 5.97 |
| Orchestrator | 5.83 |
| Build Agent | 5.71 |
| Code-Review Agent | 5.70 |
| Develop Agent | 5.65 |
| Code Agent | 5.61 |
Run history (primary model)
| Run | SoW | Date | Composite |
|---|---|---|---|
| Gemini Flash 3 (preview) | circles | Jun 2026 | 56.6 |
| Gemini Flash 3 (preview) | solarSystem | Jun 2026 | 28.9 |
| Gemini Flash 3 (preview) | circles | Jul 2026 | 53.8 |
| Gemini Flash 3 (preview) | circles | Jul 2026 | 39.1 |
| Gemini Flash 3 (preview) | circles | Jul 2026 | 56.5 |
| Gemini Flash 3 (preview) | circles | Jul 2026 | 54.4 |
Every run on Favur Evals is scored by the same deterministic engine, on the same Statements of Work. The benchmark is self-funded — no vendor sponsorship, credits, or grants.