Gemini 4 Argon (Low): benchmark results
Provider: Google. Released 2026-09-30. Access: API.
Aggregate Rating 1921 ± 26, rank #23 of 2169 rated models, from 14 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Conceptual Reasoning Index - Decision Theory (DTBench) | 96 | Chance-Corrected Score (0-100) | 97.9 |
| ARC-AGI-3 | 49.79 | Accuracy (%) | 94 |
| DuelLab Overall | 80.98 | GameBench2 normalized score (0–100) | 92.9 |
| Bug Hunt Bench - LMS | 22.7 | Planted Bugs Fixed (out of 60) | 91.5 |
| Conceptual Reasoning Index - Argument Evaluation (LMCA) | 62.92 | Chance-Corrected Score (0-100) | 90.2 |
| Conceptual Reasoning Index | 72.41 | Chance-Corrected Score (0-100) | 88.8 |
| Argo-Bench - Solved | 25.24 | Share of the 210 tasks solved (%), a task counting as solved | 87.5 |
| CursorBench 4.0 | 50.6 | Score (%) | 86.3 |
| Bug Hunt Bench | 35.3 | Planted Bugs Fixed (out of 105) | 80.6 |
| Argo-Bench | 38.81 | Mean task score (0-100) over 210 enterprise data-science tas | 80.4 |
| Conceptual Reasoning Index - Consistency (ACCoRD) | 77.3 | Chance-Corrected Score (0-100) | 77.5 |
| Bug Hunt Bench - VS Code Extension | 12.7 | Planted Bugs Fixed (out of 45) | 60.9 |
Interactive version: theaggregate.ai/model?slug=gemini-4-argon-low · How It Works · Data refreshed daily, snapshot 2026-10-23.