System record

o3 (high) / TasteVal

Organization: OpenAI · base model: o3, Opus 4.8 (fixed Coder)

System and setup

Reported system information
Model versiono3 (high)
Software used to run the modelTasteVal ReAct
Software versionarXiv:2610.06824v1
Tools available to the modelFixed Opus 4.8 Coder, two monitors and liaison; one H100; internet and experimental reports.
Release dateNot reported

Autonomy assessments

No autonomy assessment is published for this system.

Published capability results

Benchmark and indicator evidence
MeasureVersionReported resultDate
TasteValOctober 2026 paper v1 · 20-model main cohort0.08×2026-10-05
TasteValOctober 2026 paper v1 · 20-model main cohort0.43×2026-10-05

Related research

Additional system details
System ID
tasteval-o3
Organization
OpenAI
Base model
o3, Opus 4.8 (fixed Coder)
Release date
Not reported
Assessment records
0
Recursive evidence records
0