Instruction following
IFBench leaderboard
As of 2026-10-07, the highest measured score on IFBench is 83.3% by Grok 4.3 (medium).
Measured scores: artificial-analysis.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | Grok 4.3 (medium) | 83.3% | measured |
| 2 | MiniMax-M3 | 82.9% | measured |
| 3 | Grok 4.20 0309 | 82.9% | measured |
| 4 | Nemotron 3 Ultra | 81.4% | measured |
| 5 | Grok 4.3 (high) | 81.3% | measured |
| 6 | Grok 4.20 0309 v2 | 81.2% | measured |
| 7 | Grok 4.3 (low) | 81.0% | measured |
All leaderboards
Agentic · terminal
Agentic · tools
- AA-AnalystAgent
- APEX-Agents
- AutomationBench
- EnterpriseOps-Gym
- Harvey LAB
- ITBench SRE
- τ³-Bench Banking
- τ²-Bench Telecom