What it is
Published Best Overall (Humanity's Last Exam) scores.
← Back to leaderboards
BENCHMARK RECORD
Humanity's Last Exam
Source-linked result details from the TopLLM database.
BENCHMARK RECORD · 2026-09-15
Humanity's Last Exam
What it tests
Published Best Overall (Humanity's Last Exam) scores.
How to read it
Scores are benchmark-specific and missing values are not estimated.
25 published entries65.0 top reported scoreoverall evaluation domain
Published leaders
Each bar is scaled only within this benchmark.
01Claude Fable 5.165.0
02Claude Mythos 5.165.0Unresolved model record
03Claude Opus 564.7Unresolved model record
04Claude Mythos 564.505Claude Opus 4.857.906Claude Sonnet 557.4Unresolved model record
07GPT-6 Astra57.208Kimi K356.0Unresolved model record
09GLM 5.3 Flash55.310GLM 5.254.711Kimi K2.654.0Unresolved model record
12DeepSeek V4 Flash51.613DeepSeek V4 Pro48.214GPT-5.6 Sol47.2Unresolved model record
15Gemini 3 Pro45.816Kimi K2 Thinking44.917Gemini 3.1 Pro44.418GPT-5.5 Pro43.1Unresolved model record
19GPT-5.541.420Gemini 3.5 Flash40.221Claude Opus 4.640.022GPT-535.2Unresolved model record
23Kimi K2.530.1Unresolved model record
24Grok 425.425gpt-oss-120b14.9Unresolved model record
Scores are benchmark-specific. Missing values remain N/A and are never inferred from another evaluation.
Compare in the benchmark matrix →