bix-45-q1 | Current BixBench results

New token-optimization experiment

Three additional GPT-5.6 Sol Galaxy replicates are available separately. Previous results remain unchanged.

View the October 2026 batch and its run records

Question

What is the p-value from a Mann-Whitney U test comparing RCV scores between animal and fungal orthologs using PhyKIT's rcv function?

Accepted answer: 7.6968e-54 or 1.5197572608715265e-56
Domain: phylogenetics

Comparison

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 4.28M total	3/3 replicate 1 replicate 2 replicate 3 2.42M total
Codex GPT-5.6 Sol	3/3 replicate 1 replicate 2 replicate 3 5.11M total	3/3 replicate 1 replicate 2 replicate 3 3.34M total
Codex GPT-5.6 Luna	3/3 replicate 1 replicate 2 replicate 3 12.98M total	3/3 replicate 1 replicate 2 replicate 3 7.45M total
DeepSeek V4 Pro via Codex	3/3 replicate 1 replicate 2 replicate 3 6.30M total	3/3 replicate 1 replicate 2 replicate 3 16.24M total
DeepSeek V4 Pro via Claude Codesuperseded	3/3 replicate 1 replicate 2 replicate 3 4.59M total	1/3 replicate 1 replicate 2 replicate 3 1.98M total

Execution condition

Galaxy-API code with skills

Codex GPT-5.5

3 primary replicates

Codex GPT-5.5

replicate 1

passUDT + standard tools

Answer1.5197572608715265e-56

Workflow

Galaxy UDTNonparametric Rank TestsPhyKITMAFFT

Total tokens: 871.6K
Output: 7.8K

Galaxy history Run trace

Codex GPT-5.5

replicate 2

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank TestsMAFFT

Total tokens: 1.69M
Output: 10.6K

Galaxy history Run trace

Codex GPT-5.5

replicate 3

passUDT + standard tools

Answer1.5197572608715265e-56

Workflow

Galaxy UDTNonparametric Rank TestsPhyKITMAFFT

Total tokens: 1.72M
Output: 11.2K

Galaxy history Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank TestsMAFFT

Total tokens: 1.42M
Output: 7.3K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank TestsMAFFT

Total tokens: 2.15M
Output: 8.9K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 1.54M
Output: 8.3K

Galaxy history Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank TestsMAFFT

Total tokens: 3.41M
Output: 19.9K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 6.45M
Output: 26.4K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 3.11M
Output: 19.1K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 879.6K
Output: 10.8K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 1.53M
Output: 12.6K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank TestsMAFFT

Total tokens: 3.89M
Output: 24.1K

Galaxy history Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 2.08M
Output: 17.6K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 945.3K
Output: 10.8K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer1.5197572608715265e-56

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 1.57M
Output: 20.0K

Galaxy history Run trace

Execution condition

Open-ended code with skills

Codex GPT-5.5

3 primary replicates

Codex GPT-5.5

replicate 1

pass

Answer1.5197572608715265e-56

Workflow

Local analysisBUSCOPhyKITSciPyMAFFTMann-Whitney U

Total tokens: 582.4K
Output: 5.9K

Run trace

Codex GPT-5.5

replicate 2

pass

Answer1.5197572608715265e-56

Workflow

Local analysisSciPyPhyKITMAFFTMann-Whitney Umedian summary

Total tokens: 1.18M
Output: 8.7K

Run trace

Codex GPT-5.5

replicate 3

pass

Answer1.5197572608715265e-56

Workflow

Local analysisPhyKITSciPyMAFFTMann-Whitney Umedian summary

Total tokens: 653.2K
Output: 6.7K

Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer7.6967608298013025e-54

Workflow

Local analysis

Total tokens: 1.29M
Output: 9.6K

Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer7.696760829801303e-54

Workflow

Local analysis

Total tokens: 773.7K
Output: 9.4K

Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer7.696760829801303e-54

Workflow

Local analysis

Total tokens: 1.28M
Output: 9.2K

Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer7.696760829801303e-54

Workflow

Local analysis

Total tokens: 3.21M
Output: 18.9K

Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer1.5197572608715265e-56

Workflow

Local analysis

Total tokens: 1.25M
Output: 16.3K

Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer7.696760829801303e-54

Workflow

Local analysis

Total tokens: 2.99M
Output: 23.9K

Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer7.696760829801303e-54

Workflow

Local analysisPhyKITSciPyMAFFTBiopythonMann-Whitney U

Total tokens: 6.94M
Output: 33.0K

Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer7.696760829801303e-54

Workflow

Local analysisPhyKITBUSCOMAFFTSciPyMann-Whitney U

Total tokens: 4.78M
Output: 32.4K

Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer7.696760829801303e-54

Workflow

Local analysisBUSCOPhyKITMAFFTSciPyMann-Whitney U

Total tokens: 4.52M
Output: 29.1K

Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

fail

Why it missed

Version and alignment-scope mismatchThis run used newer PhyKIT RCV semantics and restricted the comparison to shared orthologs, producing p=4.0287e-55. PhyKIT 2.0.3 on all 241 animal versus 255 fungal alignments reproduces the accepted p=7.6968e-54.

Answer4.028688057945899e-55

Workflow

Local analysisBUSCOMAFFTPhyKITSciPyMann-Whitney U

Total tokens: 713.6K
Output: 7.0K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

fail

Why it missed

Answer4.028688057945899e-55

Workflow

Local analysisBUSCOPhyKITMAFFTSciPyMann-Whitney U

Total tokens: 864.5K
Output: 7.8K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer1.519757260871527e-56

Workflow

Local analysisPhyKITMAFFTSciPyMann-Whitney U

Total tokens: 405.6K
Output: 5.3K

Run trace

bix-45-q1 · October 2026 · three additional replicates

GPT-5.6 Sol · token optimization

Galaxy-API code, high reasoning, Fast service. Previous runs are retained above.

Galaxy-API code · October 2026

replicate 1

Incorrect

Answer1.5197572608715265e-56

Total tokens: 758.8K
Input: 755.3K
Cached input: 700.2K
Output: 3.4K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 2

Incorrect

Answer1.5197572608715265e-56

Total tokens: 691.9K
Input: 683.6K
Cached input: 633.3K
Output: 8.3K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 3

Incorrect

Answer1.5197572608715265e-56

Total tokens: 491.5K
Input: 487.1K
Cached input: 452.9K
Output: 4.4K

Run trace Final answer Galaxy history

Open-ended code averaged 1.11M tokens; previous website Galaxy averaged 1.70M; the new Galaxy runs averaged 0.65M, or 0.58× 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.

Optimization overview and all 150 runs