replicate 1
19159- Total tokens
- 886.8K
- Output
- 6.2K
BixBench current item
Three additional GPT-5.6 Sol Galaxy replicates are available separately. Previous results remain unchanged.
How many individual methylation measurements (rows) are removed when filtering out measurements that do not show >90% or <10% methylation in the Zebra Finch dataset?
Pass counts and token totals across the same three replicates.
| Agent | Galaxy-API code | Open-ended code |
|---|---|---|
| Codex GPT-5.5 | 3/3 replicate 1 replicate 2 replicate 3 2.22M total | 3/3 replicate 1 replicate 2 replicate 3 196.9K total |
| Codex GPT-5.6 Sol | 2/3 replicate 1 replicate 2 replicate 3 2.80M total | 3/3 replicate 1 replicate 2 replicate 3 190.4K total |
| Codex GPT-5.6 Luna | 1/3 replicate 1 replicate 2 replicate 3 13.90M total | 3/3 replicate 1 replicate 2 replicate 3 245.9K total |
| DeepSeek V4 Pro via Codex | 3/3 replicate 1 replicate 2 replicate 3 3.34M total | 3/3 replicate 1 replicate 2 replicate 3 267.8K total |
| DeepSeek V4 Pro via Claude Codesuperseded | 3/3 replicate 1 replicate 2 replicate 3 6.39M total | 2/3 replicate 1 replicate 2 replicate 3 4.37M total |
Execution condition
1915919159191591915919159191601915919160539191591915919159191591915919159Execution condition
1915919159191591915919159191591915919159191591915919159191591915919159not producedbix-52-q7 · October 2026 · three additional replicates
Galaxy-API code, high reasoning, Fast service. Previous runs are retained above.
191591915919159Open-ended code averaged 0.06M tokens; previous website Galaxy averaged 0.93M; the new Galaxy runs averaged 0.57M, or 8.93× Open-ended code. Each mean covers three replicates of this same task with GPT-5.6 Sol.
Cached input is already included in input tokens. Full nested tool/job evidence is in each HF run's galaxy_execution_records.tar.gz, with an accompanying member index.