HELM AIR-Bench 2024 - #5.15: Aircraft navigation/Air traffic control: leaderboard

Metric: Refusal Rate (%). Source: crfm.stanford.edu. 87 models tracked.

Top models

#ModelScore
1GPT-5100
2Claude 3.5 Sonnet (20241022)100
3Gemini 1.5 Flash (002)100
4Claude 3.5 Sonnet (20240620)100
5GPT-5 Nano100
6GPT-OSS-120B100
7Qwen 3 Next 80B A3B (Thinking)100
8GPT-4 Turbo100
9Claude Haiku 4.5 (20251001)100
10Llama 4 Maverick Instruct FP8100
11O4 Mini (2025-04-16)100
12O3 (2025-04-16)100
13Claude 3 Opus (20240229)100
14Claude 3 Haiku (20240307)100
15O3 Mini (2025-01-31)100

Interactive version: theaggregate.ai/benchmark?slug=helm-air-bench-2024-5-15-aircraft-navigation-air-traffic-control · How It Works · Data refreshed daily, snapshot 2026-09-19.