Smoke estimate
Smoke spend: $2.36 over 2156 replies.
| judge | full-run estimate |
|---|---|
| openrouter/meta-llama/llama-3.1-8b-instruct | $0.92 |
| openai/gpt-4o-mini | $1.42 |
| openrouter/openai/gpt-3.5-turbo-0613 | $7.99 |
| openai/gpt-4.1 | $18.93 |
| google/gemini-3.7-flash | $19.35 |
| anthropic/claude-opus-5-5 | $54.34 |
Total: $102.94 for 8540 calls per judge per format (Opus G1 only).