bix-11-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 percentage of fungal genes have treeness values above 0.06?

Accepted answer: 35%
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 4.34M total	3/3 replicate 1 replicate 2 replicate 3 955.9K total
Codex GPT-5.6 Sol	3/3 replicate 1 replicate 2 replicate 3 3.98M total	3/3 replicate 1 replicate 2 replicate 3 1.19M total
Codex GPT-5.6 Luna	3/3 replicate 1 replicate 2 replicate 3 41.09M total	3/3 replicate 1 replicate 2 replicate 3 2.32M total
DeepSeek V4 Pro via Codex	3/3 replicate 1 replicate 2 replicate 3 4.91M total	3/3 replicate 1 replicate 2 replicate 3 4.85M total
DeepSeek V4 Pro via Claude Codesuperseded	3/3 replicate 1 replicate 2 replicate 3 3.96M total	3/3 replicate 1 replicate 2 replicate 3 5.76M total

Execution condition

Galaxy-API code with skills

Codex GPT-5.5

3 primary replicates

Codex GPT-5.5

replicate 1

pass

Answer35.34136546184739%

Workflow

Phykit MetricsBUSCO

Total tokens: 1.19M
Output: 7.0K

Galaxy history Run trace

Codex GPT-5.5

replicate 2

pass

Answer35.341365%

Workflow

Phykit MetricsMAFFTIQ-TREE

Total tokens: 1.73M
Output: 7.2K

Galaxy history Run trace

Codex GPT-5.5

replicate 3

passUDT-only

Answer35.3413654618%

Workflow

Galaxy UDTIQ-TREEPhyKITBiopython

Total tokens: 1.42M
Output: 15.2K

Galaxy history Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer35.34%

Workflow

Phykit Metrics

Total tokens: 1.15M
Output: 5.0K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer35.34%

Workflow

Phykit Metrics

Total tokens: 891.4K
Output: 5.2K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer35.34%

Workflow

Phykit MetricsIQ-TREEMAFFT

Total tokens: 1.94M
Output: 6.7K

Galaxy history Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer35.341365%

Workflow

Phykit MetricsTable ComputeFilter Tabularmore stepsIQ-TREEMAFFT

Total tokens: 26.22M
Output: 44.8K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer35.34%

Workflow

UnzipPhykit MetricsIQ-TREEMAFFT

Total tokens: 4.95M
Output: 14.9K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 3

passUDT + standard tools

Answer35.341365461847%

Workflow

UnzipConcatenate datasetsCollapse Datasetmore stepsMAFFTIQ-TREE

Total tokens: 9.92M
Output: 31.7K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer35.341365461847%

Workflow

Phykit Metrics

Total tokens: 982.7K
Output: 9.3K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer35.34%

Workflow

Phykit Metrics

Total tokens: 888.6K
Output: 6.6K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer35%

Workflow

Phykit MetricsBUSCOMAFFTmean summarymedian summary

Total tokens: 3.04M
Output: 20.4K

Galaxy history Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer35.34%

Workflow

Phykit MetricsIQ-TREEMAFFT

Total tokens: 2.08M
Output: 22.1K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

pass

Answer35.34%

Workflow

Phykit Metrics

Total tokens: 1.06M
Output: 10.6K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer34.51%

Workflow

Phykit Metrics

Total tokens: 820.6K
Output: 9.0K

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

Answer35.34136546184739%

Workflow

Local analysisBUSCOIQ-TREEMAFFTmean summary

Total tokens: 426.2K
Output: 4.4K

Run trace

Codex GPT-5.5

replicate 2

pass

Answer35.34136546184739%

Workflow

Local analysisBUSCOMAFFTIQ-TREENewick analysis

Total tokens: 269.1K
Output: 4.3K

Run trace

Codex GPT-5.5

replicate 3

pass

Answer35.34136546184739%

Workflow

Local analysisMAFFTIQ-TREENewick analysis

Total tokens: 260.6K
Output: 4.9K

Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer35.34%

Workflow

Local analysis

Total tokens: 546.7K
Output: 5.1K

Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer35.34136546184739%

Workflow

Local analysis

Total tokens: 290.6K
Output: 5.0K

Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer35.34%

Workflow

Local analysis

Total tokens: 348.1K
Output: 5.3K

Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer35.34136546184739%

Workflow

Local analysis

Total tokens: 608.5K
Output: 12.8K

Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer35.34%

Workflow

Local analysis

Total tokens: 1.09M
Output: 12.2K

Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer35.34%

Workflow

Local analysis

Total tokens: 615.5K
Output: 6.8K

Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer35.34%

Workflow

Local analysisBUSCOMAFFTPhyKITNewick analysismedian summary

Total tokens: 1.51M
Output: 17.7K

Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer35%

Workflow

Local analysisBUSCOPhyKITMAFFTNewick analysismean summary

Total tokens: 2.18M
Output: 15.2K

Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer35%

Workflow

Local analysisMAFFTIQ-TREEPhyKIT

Total tokens: 1.16M
Output: 10.3K

Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer34.51%

Workflow

Local analysisBUSCOMAFFTIQ-TREEmean summarymedian summary

Total tokens: 2.50M
Output: 41.9K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

pass

Answer35.34%

Workflow

Local analysisBUSCOMAFFTIQ-TREENewick analysis

Total tokens: 523.5K
Output: 8.0K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer34.5098%

Workflow

Local analysisMAFFTBiopythonIQ-TREEBUSCO

Total tokens: 2.74M
Output: 34.7K

Run trace

bix-11-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

Answer35.34%

Total tokens: 717.2K
Input: 711.3K
Cached input: 673.4K
Output: 5.8K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 2

Correct

Answer35.34%

Total tokens: 505.5K
Input: 502.0K
Cached input: 462.6K
Output: 3.5K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 3

Correct

Answer35.34%

Total tokens: 574.2K
Input: 570.4K
Cached input: 524.8K
Output: 3.8K

Run trace Final answer Galaxy history

Open-ended code averaged 0.40M tokens; previous website Galaxy averaged 1.33M; the new Galaxy runs averaged 0.60M, or 1.52× 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