Skip to content
The Aggregate

Unified LLM rankings, daily

Models What's New Trends Benchmarks Skill Maps Task Explorer
How It Works Guesswork Metrics
The Aggregate
Loading data...

OneMillion-Bench — leaderboard

Metric: Pass Rate (%). Source: humanlaya.com. 36 models tracked.

Top models

#ModelScore
1Claude Opus 4.643.5
2GPT-5.232
3GPT-5.431.5
4GPT-5.3 Codex29
5Gemini 3 Pro (Preview)28.5
6Qwen 3.5 Plus22.5
7Grok 418.5
8Kimi K2.517.5
9O317
10DeepSeek V3.2 Speciale16
11Seed 2.0 Pro13.5
12Qwen 3 Max12
13O4 Mini10.5
14GLM-59.5
15MiniMax-M2.18

Interactive version: theaggregate.ai/benchmark?slug=onemillion-bench · How It Works · Data refreshed daily, snapshot 2026-07-27.

Built by Mikhail Doroshenko — AI researcher, co-author of Humanity’s Last Exam. Independent project; no affiliation with any AI lab.