bix-11-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 difference between median treeness values for fungi versus animals? Report the difference as a decimal proportion (not percentage).

Accepted answer: 0.05
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 2.96M total	3/3 replicate 1 replicate 2 replicate 3 988.4K total
Codex GPT-5.6 Sol	3/3 replicate 1 replicate 2 replicate 3 5.24M total	3/3 replicate 1 replicate 2 replicate 3 840.5K total
Codex GPT-5.6 Luna	3/3 replicate 1 replicate 2 replicate 3 53.13M total	3/3 replicate 1 replicate 2 replicate 3 2.11M total
DeepSeek V4 Pro via Codex	3/3 replicate 1 replicate 2 replicate 3 7.94M total	3/3 replicate 1 replicate 2 replicate 3 4.44M total
DeepSeek V4 Pro via Claude Codesuperseded	3/3 replicate 1 replicate 2 replicate 3 6.71M total	3/3 replicate 1 replicate 2 replicate 3 4.96M total

Execution condition

Galaxy-API code with skills

Codex GPT-5.5

3 primary replicates

Codex GPT-5.5

replicate 1

passUDT + standard tools

Answer0.050099999999999999

Workflow

Phykit MetricsGalaxy UDTBUSCOmedian summary

Total tokens: 1.18M
Output: 8.5K

Galaxy history Run trace

Codex GPT-5.5

replicate 2

pass

Answer0.0501

Workflow

Phykit Metricsmedian summary

Total tokens: 1.22M
Output: 7.3K

Galaxy history Run trace

Codex GPT-5.5

replicate 3

passUDT-only

Answer0.050139847768049167

Workflow

Galaxy UDTIQ-TREEMAFFTPhyKITmedian summary

Total tokens: 569.4K
Output: 9.4K

Galaxy history Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer0.0501

Workflow

Phykit Metricsmedian summary

Total tokens: 1.58M
Output: 8.0K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer0.0501

Workflow

Phykit MetricsDatamash summary

Total tokens: 2.10M
Output: 9.8K

Galaxy history Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer0.0501

Workflow

Phykit Metricsmedian summary

Total tokens: 1.56M
Output: 7.5K

Galaxy history Run trace

History isolation noteReplicate 2 created this history first. During recovery, replicate 3 selected it by name, copied fresh seed inputs, and submitted independent PhyKIT jobs. This is a history-isolation violation without detected output reuse.

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer0.0501

Workflow

Phykit MetricsBUSCOmedian summary

Total tokens: 2.73M
Output: 18.6K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 2

passUDT + standard tools

Answer0.0501

Workflow

UnzipFilter From FileGalaxy UDTmore stepsPhyKITmedian summary

Total tokens: 45.64M
Output: 56.8K

Galaxy history Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer0.0501

Workflow

Phykit MetricsDatamash summaryBUSCO

Total tokens: 4.76M
Output: 21.1K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer0.0501

Workflow

Phykit MetricsSummary statistics

Total tokens: 3.69M
Output: 26.2K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer0.0501

Workflow

Phykit Metrics

Total tokens: 722.7K
Output: 8.9K

Galaxy history Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer0.0501

Workflow

Phykit Metricsmedian summary

Total tokens: 3.52M
Output: 24.8K

Galaxy history Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer0.0501

Workflow

Phykit MetricsIQ-TREEBUSCOmedian summary

Total tokens: 1.63M
Output: 17.4K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

pass

Answer0.0501

Workflow

Phykit Metricsmedian summary

Total tokens: 3.75M
Output: 31.5K

Galaxy history Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

passUDT-only

Answer0.0501

Workflow

Galaxy UDTPhyKITmedian summary

Total tokens: 1.33M
Output: 17.7K

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

Answer0.05013984776804917

Workflow

Local analysisBUSCOIQ-TREEBiopythonNewick analysismedian summary

Total tokens: 268.2K
Output: 5.9K

Run trace

Codex GPT-5.5

replicate 2

pass

Answer0.05013984776804917

Workflow

Local analysisBUSCOIQ-TREEMAFFTmedian summary

Total tokens: 421.7K
Output: 5.3K

Run trace

Codex GPT-5.5

replicate 3

pass

Answer0.05013984776804917

Workflow

Local analysisMAFFTIQ-TREEmedian summaryNewick analysis

Total tokens: 298.5K
Output: 4.8K

Run trace

Codex GPT-5.6 Sol

3 primary replicates

Codex GPT-5.6 Sol

replicate 1

pass

Answer0.05013984776804917

Workflow

Local analysis

Total tokens: 255.1K
Output: 4.3K

Run trace

Codex GPT-5.6 Sol

replicate 2

pass

Answer0.05013984776804917

Workflow

Local analysis

Total tokens: 302.4K
Output: 5.2K

Run trace

Codex GPT-5.6 Sol

replicate 3

pass

Answer0.0501398477680492

Workflow

Local analysis

Total tokens: 283.0K
Output: 6.3K

Run trace

Codex GPT-5.6 Luna

3 primary replicates

Codex GPT-5.6 Luna

replicate 1

pass

Answer0.05013984776804917

Workflow

Local analysis

Total tokens: 833.6K
Output: 13.9K

Run trace

Codex GPT-5.6 Luna

replicate 2

pass

Answer0.05014

Workflow

Local analysis

Total tokens: 736.5K
Output: 11.4K

Run trace

Codex GPT-5.6 Luna

replicate 3

pass

Answer0.05014

Workflow

Local analysis

Total tokens: 539.6K
Output: 10.6K

Run trace

DeepSeek V4 Pro via Codex

3 primary replicates

DeepSeek V4 Pro via Codex

replicate 1

pass

Answer0.0501

Workflow

Local analysisBUSCOIQ-TREEPhyKITSciPymedian summary

Total tokens: 1.45M
Output: 15.2K

Run trace

DeepSeek V4 Pro via Codex

replicate 2

pass

Answer0.05014

Workflow

Local analysisMAFFTPhyKITBiopythonNewick analysismedian summary

Total tokens: 1.04M
Output: 9.9K

Run trace

DeepSeek V4 Pro via Codex

replicate 3

pass

Answer0.05014

Workflow

Local analysisIQ-TREEMAFFTNewick analysismedian summarymean summary

Total tokens: 1.94M
Output: 15.8K

Run trace

DeepSeek V4 Pro via Claude CodeSuperseded

3 primary replicates

DeepSeek V4 Pro via Claude Code

replicate 1

pass

Answer0.0501

Workflow

Local analysisBUSCOMAFFTIQ-TREEPhyKITmedian summary

Total tokens: 1.64M
Output: 18.1K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 2

pass

Answer0.05012776191191943

Workflow

Local analysisBUSCOMAFFTIQ-TREEmedian summarymean summary

Total tokens: 813.9K
Output: 9.7K

Run trace

DeepSeek V4 Pro via Claude Code

replicate 3

pass

Answer0.05014

Workflow

Local analysisBUSCOMAFFTNewick analysismedian summarymean summary

Total tokens: 2.51M
Output: 36.1K

Run trace

bix-11-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

Answer0.0501

Total tokens: 584.5K
Input: 579.8K
Cached input: 542.7K
Output: 4.8K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 2

Correct

Answer0.0501

Total tokens: 772.5K
Input: 766.7K
Cached input: 723.6K
Output: 5.8K

Run trace Final answer Galaxy history

Galaxy-API code · October 2026

replicate 3

Correct

Answer0.0501

Total tokens: 629.1K
Input: 624.7K
Cached input: 560.9K
Output: 4.4K

Run trace Final answer Galaxy history

Open-ended code averaged 0.28M tokens; previous website Galaxy averaged 1.75M; the new Galaxy runs averaged 0.66M, or 2.36× 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