Opus 5 Takes the Top Spot in Intelligence Index, Beating Fable 5 at Half the Cost
Artificial Analysis scores Claude Opus 5 at 61 on its Intelligence Index, the highest result recorded. Three variants of the model land in the top five. Here is the full picture, including which numbers come from Anthropic itself.
Claude Opus 5 entered the independent Artificial Analysis ranking at 61 points on the Intelligence Index, the highest score on the board. It is the first model to pass Fable 5, and it does so at half the price per token.
The current top five
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 5 (max effort) | 61 |
| 2 | Claude Opus 5 (xhigh effort) | 60 |
| 3 | Claude Fable 5 (max effort) | 60 |
| 4 | GPT-5.6 Sol (max) | 59 |
| 5 | Claude Opus 5 (high effort) | 59 |
Read that table twice. Opus 5 on high effort ties GPT-5.6 Sol on max effort. One model now competes in several price brackets at once, depending on where you set the dial.
Vendor benchmarks, labelled as such
The following results come from Anthropic’s own launch material. Treat them as vendor claims rather than independent measurement, but they are specific enough to be checked:
- Frontier-Bench v0.1 – surpasses all other models and more than doubles Opus 4.8’s performance at a lower cost per task.
- ARC-AGI 3 – three times the score of the next-best model on abstract reasoning.
- CursorBench 3.2 – within 0.5% of Fable 5’s peak score at max effort, at half the cost per task.
- OSWorld 2.0 – beats Fable 5’s best computer-use result at just over a third of the cost.
- Zapier AutomationBench – roughly 1.5x the pass rate of the next-best model at the same cost per task, and it clears more tasks than rivals even on the lowest effort setting.
- Organic chemistry – 10.2 percentage points above Opus 4.8. Protein tasks – 7.7 points higher.
What is still missing
Opus 5 has not appeared in Text Arena yet, where rankings are built from millions of blind human comparisons over weeks. Fable 5 still leads there at 1507 Elo, with Anthropic models occupying most of the top six. Measured throughput and time to first token for Opus 5 are also not published yet.
So the honest summary is this: on aggregated benchmarks Opus 5 is the strongest model available today, at a price that undercuts the model it just passed. On human preference, the jury has not convened.
