# FrontierMath (Tier 4) — AI model rankings

Very hard, research-level math problems. Tier 4 is the hardest — close to what professional research mathematicians tackle. Higher is better.

6 tracked models have a published FrontierMath (Tier 4) score. Higher is better. Scores come from published lab reports and benchmark sources; recorded sources appear beside the scores.

## Ranking

| Rank | Model | Developer | Score | Source | Released |
| --- | --- | --- | --- | --- | --- |
| 1 | GPT-5.5-Pro | OpenAI | 39.6% | Lab | Apr 23 2026 |
| 2 | GPT-5.4-Pro | OpenAI | 38% | Lab | Mar 5 2026 |
| 3 | GPT-5.5 | OpenAI | 35.4% | Lab | Apr 23 2026 |
| 4 | GPT-5.4 | OpenAI | 27.1% | Lab | Mar 5 2026 |
| 5 | Claude Opus 4.7 | Anthropic | 22.9% | Lab | Apr 16 2026 |
| 6 | Gemini 3.1 Pro | Google | 16.7% | Lab | Feb 19 2026 |


---

Canonical page: https://aireleasetracker.com/benchmark/frontiermath-t4
Full dataset: https://aireleasetracker.com/llms-full.txt · JSON: https://aireleasetracker.com/models.json
Source: AI Release Tracker (https://aireleasetracker.com). Most benchmark scores come from lab launch material; gathered results identify the leaderboard that published them.