bix-38-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 fold-change in median tree length between fungi and animals after calculating tree lengths with PhyKIT?

Accepted answer: 2x larger in fungi
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 3.42M total	3/3 replicate 1 replicate 2 replicate 3 1.42M total
Codex GPT-5.6 Sol	3/3 replicate 1 replicate 2 replicate 3 4.94M total	3/3 replicate 1 replicate 2 replicate 3 1.45M total
Codex GPT-5.6 Luna	3/3 replicate 1 replicate 2 replicate 3 11.52M total	3/3 replicate 1 replicate 2 replicate 3 2.85M total
DeepSeek V4 Pro via Codex	3/3 replicate 1 replicate 2 replicate 3 3.50M total	3/3 replicate 1 replicate 2 replicate 3 5.39M total
DeepSeek V4 Pro via Claude Codesuperseded	3/3 replicate 1 replicate 2 replicate 3 3.97M total	3/3 replicate 1 replicate 2 replicate 3 2.88M total

Execution condition

Galaxy-API code with skills

Codex GPT-5.5

3 primary replicates

Codex GPT-5.5

replicate 1

pass

Answer2.177508966875307686897812786

Workflow

Phykit MetricsBUSCOmedian summary

Total tokens: 1.38M
Output: 8.0K

Galaxy history Run trace

Codex GPT-5.5

replicate 2

pass

Answer2.177508966875308

Workflow

Phykit Metricsmedian summary

Total tokens: 986.7K
Output: 8.1K

Galaxy history Run trace

Codex GPT-5.5

replicate 3

pass

Answer2.17750896688

Workflow

Phykit Metricsmedian summary

Total tokens: 1.05M
Output: 8.7K

Galaxy history Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer2.18

Workflow

Phykit Metricsmedian summary

Total tokens: 1.61M
Output: 8.5K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer2.1775

Workflow

Phykit Metricsmedian summary

Total tokens: 1.55M
Output: 7.4K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer2.177508966875308

Workflow

Phykit Metricsmedian summary

Total tokens: 1.78M
Output: 8.2K

Galaxy history Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer2.1775

Workflow

Phykit MetricsDatamash summary

Total tokens: 5.82M
Output: 20.7K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer2.1775089668753077

Workflow

Phykit MetricsBUSCOmedian summary

Total tokens: 2.97M
Output: 15.4K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer2.1775089668753077

Workflow

Phykit Metricsmedian summary

Total tokens: 2.72M
Output: 18.7K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer2.1775

Workflow

Phykit Metricsmedian summarymean summary

Total tokens: 2.06M
Output: 21.7K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer2.177508967

Workflow

Phykit Metricsmedian summary

Total tokens: 679.1K
Output: 7.6K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer2.17750896687531

Workflow

Phykit Metrics

Total tokens: 758.8K
Output: 11.1K

Galaxy history Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer2.1775

Workflow

Phykit Metrics

Total tokens: 1.09M
Output: 12.2K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

pass

Answer2.1775

Workflow

Phykit Metricsmedian summaryfold-change calculation

Total tokens: 953.5K
Output: 14.9K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer2.177509

Workflow

Phykit Metricsmedian summaryfold-change calculationmean summary

Total tokens: 1.92M
Output: 17.6K

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

Answer2.1775089668753077

Workflow

Local analysisPhyKITNewick analysis

Total tokens: 400.0K
Output: 4.5K

Run trace

Codex GPT-5.5

replicate 2

pass

Answer2.177508966875308

Workflow

Local analysisPhyKITmedian summary

Total tokens: 453.3K
Output: 4.2K

Run trace

Codex GPT-5.5

replicate 3

pass

Answer2.17750896688

Workflow

Local analysisPhyKITMAFFTmedian summary

Total tokens: 563.8K
Output: 5.3K

Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer2.1775089668753077

Workflow

Local analysis

Total tokens: 484.0K
Output: 5.9K

Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer2.18

Workflow

Local analysis

Total tokens: 326.8K
Output: 3.9K

Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer2.18

Workflow

Local analysis

Total tokens: 641.9K
Output: 5.9K

Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer2.1775089668753078

Workflow

Local analysis

Total tokens: 879.5K
Output: 8.8K

Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer2.177508966875308

Workflow

Local analysis

Total tokens: 1.27M
Output: 8.8K

Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer2.177508966875308

Workflow

Local analysis

Total tokens: 693.8K
Output: 9.9K

Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer2.177508966875308

Workflow

Local analysisBUSCOIQ-TREEMAFFTPhyKITmedian summary

Total tokens: 1.26M
Output: 15.1K

Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer2.1775

Workflow

Local analysisIQ-TREEPhyKITBiopythonSciPymedian summary

Total tokens: 1.54M
Output: 12.8K

Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer2.177509

Workflow

Local analysisBUSCOIQ-TREEMAFFTPhyKITNewick analysis

Total tokens: 2.59M
Output: 20.4K

Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer2.177509

Workflow

Local analysisBUSCOPhyKITmedian summaryNewick analysisMAFFT

Total tokens: 1.01M
Output: 8.2K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

pass

Answer2.177509

Workflow

Local analysisPhyKITBUSCOMAFFTmedian summaryfold-change calculation

Total tokens: 733.8K
Output: 7.2K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer2.177508966875308

Workflow

Local analysisBUSCOPhyKITMAFFTmedian summaryNewick analysis

Total tokens: 1.13M
Output: 10.4K

Run trace

bix-38-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

Correct

Answer2.17750896687531

Total tokens: 486.4K
Input: 482.6K
Cached input: 449.7K
Output: 3.8K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 2

Correct

Answer2.17750896687531

Total tokens: 529.8K
Input: 524.8K
Cached input: 488.6K
Output: 4.9K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 3

Correct

Answer2.1775089668753078

Total tokens: 499.0K
Input: 495.1K
Cached input: 459.4K
Output: 3.9K

Run trace Final answer Galaxy history

Open-ended code averaged 0.48M tokens; previous website Galaxy averaged 1.65M; the new Galaxy runs averaged 0.51M, or 1.04× 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