The 11 Aug 2026 Artificial Analysis snapshot — the file we ingest into /leaderboard — has Anthropic in the first four slots. OpenAI's first row is GPT-5.6 Sol (max) at rank 5.
| Rank | Model | Company | Score |
|---|---|---|---|
| 1 | Claude Opus 5 (Adaptive Reasoning, Max Effort) | Anthropic | 63.1 |
| 2 | Claude Opus 5 (Adaptive Reasoning, Xhigh Effort) | Anthropic | 62.5 |
| 3 | Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | Anthropic | 62.1 |
| 4 | Claude Opus 5 (Adaptive Reasoning, High Effort) | Anthropic | 61.5 |
| 5 | GPT-5.6 Sol (max) | OpenAI | 60.9 |
An earlier version of this page had Sol at #4 with a score of 59. That was a previous snapshot. The table above matches content/leaderboard_data.json dated 11 Aug 2026.
How to read it
Opus 5 Max Effort still leads the general intelligence index. Sol (max) is 2.2 points back. Sol (xhigh) is rank 7 at 59. We don't run these evals. Artificial Analysis does.
Pick from the board, not from a vibe:
- Mixed research, coding, and writing: start with Opus 5 Max Effort.
- Already on Sol: this snapshot does not show it overtaking Opus on the index.
- Price: this file has no list rates. Check the vendor. We don't invent $/MTok here.
What we don't claim
If a number isn't on /leaderboard or in this table, it doesn't belong in the verdict. When the edition stamp moves, re-check the live board. How we ingest: /methodology.