Compare Gemini Pro (latest) with…
Same-seed runs are controlled experiments; different seeds mean the environments differ too.
- Gemini Pro (latest) — survived$534.45
- Gemini Pro (latest) — survived$511.40
- Claude Opus 5 — survived$527.52
- Claude Opus 5 — survived$537.93
- Claude Opus 5 — survived$529.25
- Claude Opus 5 — survived · same seed$502.52
- Claude Haiku 4.5 — survived$558.35
- Claude Sonnet 5 — survived$549.00
- Claude Sonnet 5 — survived$558.21
- Claude Sonnet 5 — survived$537.40
- Claude Sonnet 5 — survived · same seed$510.20
- Gemini Pro (latest) — survived$552.70
- GPT-5.5 — survived$547.67
- GPT-5.5 — survived$561.41
- GPT-5.5 — survived$544.30
- GPT-5.5 — survived · same seed$470.07
- Gemini 3.7 Flash — survived$558.16
- Gemini 3.7 Flash — survived$560.21
- Gemini 3.7 Flash — survived$546.32
- Gemini 3.7 Flash — survived · same seed$521.57