DeepSeek V4.1 Flash (Venice) vs Gemini 3.8 Flash
On the 11 use cases both models ran, DeepSeek V4.1 Flash (Venice) won 11 to 0 (better outcome first, then lower cost).
Pass rate: DeepSeek V4.1 Flash (Venice) 100% (11 runs) vs Gemini 3.8 Flash 100% (11 runs). Average cost per run: $0.01 vs $0.31.
| Metric | DeepSeek V4.1 Flash (Venice) | Gemini 3.8 Flash |
|---|---|---|
| Runs on shared use cases | 11 | 11 |
| Pass rate | 100% (11/11) | 100% (11/11) |
| Avg cost per run | $0.01 | $0.31 |
| Avg wall time | 53s | 3m |
| Step-ins per run | 0.00 | 0.00 |
| Use cases won | 11 | 0 |
Use case by use case
- AIS collision risk: closest point of approach and alert levels: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Anchor drag alarm from a GPS track: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Best fishing windows from tide, solunar and weather data: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Best sailing angle and leg times from a polar table: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Boat fuel range with a reserve over a planned trip: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Catch log checked against a made-up fishing regulations table: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Dead reckoning, estimated position with current, and computed set and drift: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Diagnose why GitHub Actions stopped running: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Fix three bugs in a Python repo so its tests pass: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Go/no-go departure time from a gridded wind file: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
- Tide height by rule of twelfths and cosine, with under-keel clearance windows: DeepSeek V4.1 Flash (Venice) better (pass 100% vs 100%)
Only use cases both models ran are compared; all settings of each model are grouped. How we score.