bix-35-q2 | 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 Mann-Whitney U statistic value for the comparison of evolutionary rates between animals and fungi across all genes? Use PhyKIT's evo_rate function to calculate evolutionary rates.

Accepted answer: 3661
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 4.63M total	3/3 replicate 1 replicate 2 replicate 3 1.20M total
Codex GPT-5.6 Sol	3/3 replicate 1 replicate 2 replicate 3 7.03M total	3/3 replicate 1 replicate 2 replicate 3 1.77M total
Codex GPT-5.6 Luna	3/3 replicate 1 replicate 2 replicate 3 12.99M total	3/3 replicate 1 replicate 2 replicate 3 3.17M total
DeepSeek V4 Pro via Codex	3/3 replicate 1 replicate 2 replicate 3 9.72M total	3/3 replicate 1 replicate 2 replicate 3 6.21M total
DeepSeek V4 Pro via Claude Codesuperseded	2/3 replicate 1 replicate 2 replicate 3 8.39M total	1/3 replicate 1 replicate 2 replicate 3 4.49M total

Execution condition

Galaxy-API code with skills

Codex GPT-5.5

3 primary replicates

Codex GPT-5.5

replicate 1

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 1.80M
Output: 8.5K

Galaxy history Run trace

Codex GPT-5.5

replicate 2

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank TestsMAFFT

Total tokens: 1.72M
Output: 10.2K

Galaxy history Run trace

Codex GPT-5.5

replicate 3

passUDT-only

Answer3661

Workflow

Galaxy UDTMAFFTPhyKITSciPyMann-Whitney U

Total tokens: 1.11M
Output: 11.0K

Galaxy history Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank TestsMAFFTIQ-TREE

Total tokens: 3.74M
Output: 13.6K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 1.48M
Output: 5.5K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank TestsBUSCO

Total tokens: 1.81M
Output: 8.1K

Galaxy history Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank Tests

Total tokens: 4.12M
Output: 22.5K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer3661

Workflow

Phykit MetricsPhykit Tree BasedNonparametric Rank TestsMAFFTIQ-TREE

Total tokens: 3.83M
Output: 24.0K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank TestsBUSCOMAFFT

Total tokens: 5.04M
Output: 25.3K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank TestsBUSCOIQ-TREEMAFFTMann-Whitney U

Total tokens: 3.93M
Output: 28.1K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank TestsMann-Whitney U

Total tokens: 2.93M
Output: 17.6K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer3661

Workflow

Phykit MetricsNonparametric Rank TestsMAFFTIQ-TREE

Total tokens: 2.86M
Output: 17.5K

Galaxy history Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer3661

Workflow

Phykit MetricsUnzipSciPyMann-Whitney U

Total tokens: 2.93M
Output: 29.1K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

fail

Why it missed

Data scope changedThis run kept only 96 gene IDs shared by the animal and fungal archives and returned U=1820.5. Using all 100 animal and 249 fungal trees gives U=3661.

Answer1820.5

Workflow

UnzipUpload datasetPhykit Metricsmore stepsSciPyMann-Whitney U

Total tokens: 2.94M
Output: 36.0K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer3661

Workflow

Phykit MetricsSciPyMann-Whitney U

Total tokens: 2.51M
Output: 28.8K

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

Answer3661.0

Workflow

Local analysisBUSCOPhyKITMAFFTIQ-TREESciPy

Total tokens: 314.4K
Output: 7.2K

Run trace

Codex GPT-5.5

replicate 2

pass

Answer3661.0

Workflow

Local analysisPhyKITIQ-TREEMAFFTSciPyMann-Whitney U

Total tokens: 498.9K
Output: 6.3K

Run trace

Codex GPT-5.5

replicate 3

pass

Answer3661.0

Workflow

Local analysisPhyKITSciPyMAFFTIQ-TREEMann-Whitney U

Total tokens: 385.4K
Output: 7.0K

Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer3661.0

Workflow

Local analysis

Total tokens: 513.5K
Output: 4.3K

Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer3661.0

Workflow

Local analysis

Total tokens: 644.2K
Output: 5.6K

Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer3661

Workflow

Local analysis

Total tokens: 616.4K
Output: 6.0K

Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer3661.0

Workflow

Local analysis

Total tokens: 901.8K
Output: 12.3K

Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer3661

Workflow

Local analysis

Total tokens: 1.35M
Output: 13.1K

Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer3661

Workflow

Local analysis

Total tokens: 913.8K
Output: 12.8K

Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer3661

Workflow

Local analysisPhyKITSciPyIQ-TREEMAFFTMann-Whitney U

Total tokens: 2.10M
Output: 18.6K

Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer3661

Workflow

Local analysisPhyKITIQ-TREEMAFFTBUSCOSciPy

Total tokens: 2.54M
Output: 19.4K

Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer3661

Workflow

Local analysisMAFFTIQ-TREEPhyKITSciPyBiopython

Total tokens: 1.57M
Output: 11.2K

Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

fail

Why it missed

Data scope changedThis run compared only 96 genes shared by the animal and fungal archives, producing U=1820.5 instead of the all-tree result U=3661.

Answer1820.5

Workflow

Local analysisPhyKITMAFFTSciPyMann-Whitney U

Total tokens: 483.6K
Output: 6.2K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

fail

Why it missed

Data scope changedThis run narrowed the analysis to 19 genes complete across all eight species, producing U=83 instead of the all-tree result U=3661.

Answer83.0

Workflow

Local analysisPhyKITBUSCOMAFFTFastTreeBiopython

Total tokens: 2.89M
Output: 33.4K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer3661.0

Workflow

Local analysisBUSCOPhyKITIQ-TREEMAFFTSciPy

Total tokens: 1.12M
Output: 12.8K

Run trace

bix-35-q2 · 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

Correct

Answer3661

Total tokens: 575.0K
Input: 571.1K
Cached input: 528.5K
Output: 4.0K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 2

Correct

Answer3661

Total tokens: 625.4K
Input: 621.4K
Cached input: 576.1K
Output: 4.0K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 3

Correct

Answer3661

Total tokens: 567.0K
Input: 563.5K
Cached input: 532.1K
Output: 3.5K

Run trace Final answer Galaxy history

Open-ended code averaged 0.59M tokens; previous website Galaxy averaged 2.34M; the new Galaxy runs averaged 0.59M, or 1.00× 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