LLM Stats (Internal API instruction following (hard)): leaderboard
Hard subset of internal API instruction-following evaluation testing strict adherence to complex, constrained API usage instructions.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around January 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 64 | #98 |
| 2 | GPT-4.5 | 54 | #274 |
| 3 | O3 Mini | 50 | #272 |
| 4 | GPT-4.1 | 49.1 | #248 |
| 5 | GPT-4.1 Mini | 45.1 | #355 |
| 6 | GPT-4.1 Nano | 31.6 | #731 |
| 7 | GPT-4o | 29.2 | #343 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-internal-api-instruction-following-hard · How It Works · Data refreshed daily, snapshot 2026-10-20.