Editing Openai/6933ffb6-9ed8-8010-99bb-a5bc968f2d9b (section)

==== LLM-stats does provide leaderboards for many of the benchmarks you asked about. For some benchmarks, I can list the top models (often top-3 or top-5) based on their published scores. But in many cases results are “self-reported” (i.e. submitted by the model vendors or community) — which reduces how confident we can be in them. (LLM-stats itself notes that many are not “verified.”) LLM Stats<ref>{{cite web|title=LLM Stats|url=https://llm-stats.com/benchmarks/gpqa|publisher=LLM Stats|access-date=2025-12-07}}</ref> ====

Here are some of the leaderboards I could confidently access, with top models:

|     |     |
| --- | --- |
|     |     |
|     |     |
|     |     |

Additionally:
* LLM-stats lists LiveCodeBench among its benchmarks. LLM Stats<ref>{{cite web|title=LLM Stats|url=https://llm-stats.com/benchmarks|publisher=LLM Stats|access-date=2025-12-07}}</ref>
* LLM-stats also enumerates MATH 500 (classic MATH benchmark) among its math benchmarks. LLM Stats<ref>{{cite web|title=LLM Stats|url=https://llm-stats.com/benchmarks|publisher=LLM Stats|access-date=2025-12-07}}</ref>