Hurrey Inc.
GAIA-Standard Benchmark Report
2025 Edition | Evaluating Decision-Making and Reasoning
Executive Summary
Hadron (Wiseman) leads the judged cohort on situational understanding, stepwise planning, and evaluation-feedback loops. GPT-family models follow closely. Search-centric models provide strong structured retrieval.
Methodology
Prompt Cohorts
45 decision-oriented tasks (strategic, analytical, evaluative).
Evaluation Scale
0-10 (10 represents exemplary human-level output).
Model
GAIA composite scoring (human + LLM consensus).
Core Metrics
Relevance
Degree to which response addresses user intent.
Accuracy
Factual and logical correctness.
Comprehensiveness
Coverage of relevant sub-points.
Clarity
Coherence and structure of output.
SAI
Situation Awareness Index — recognition of context and constraints.
SSI
Strategic Sequencing Index — logical ordering of subtasks.
EAQ
Evaluation & Adaptation Quotient — evidence of self-assessment.
Performance Analysis
Decision-Intelligence metrics comparison across leading models.
The values in this chart are listed in the Quantitative Scores table below.
Quantitative Scores
| Model | Relevance | Accuracy | Clarity | SAI | SSI | EAQ | Avg |
|---|---|---|---|---|---|---|---|
| Hadron (Wiseman) | 9 | 9 | 9 | 9.5 | 9.2 | 9 | 8.9 |
| GPT-4.1 | 8.9 | 9 | 9 | 9 | 9 | 8.5 | 8.7 |
| GPT-5 Mini | 8.8 | 8 | 9 | 8.8 | 8.9 | 8.7 | 8.6 |
| Gemini 2.5 Pro | 8.6 | 8 | 8 | 8 | 7.8 | 7.8 | 8 |
| Grok-4 | 8.2 | 8 | 8 | 7.8 | 7.5 | 7.3 | 7.7 |
| o3-pro | 8 | 8 | 8 | 7.5 | 7.4 | 7.2 | 7.7 |
| Claude | 7.6 | 7 | 7 | 6.5 | 6.2 | 6 | 6.6 |
Recommendations
Recommended pipeline: Use Hadron/Wiseman for strategy and sequencing. Pair with Claude/o3-pro for code generation. For reproducible benchmarking, run n≥30 prompts.