Claude Opus 4.8
Anthropic · released May 28, 2026
| size | large, our estimate, approximate; not disclosed by the manufacturer |
| context window | 1,000,000 tokens |
| longest response | 128,000 tokens |
| knowledge cutoff | January 2026 |
| open weights | no |
| API price | $5 per million input tokens, $25 per million output tokens |
| input and output | text and images in, text out |
Manufacturer data (announcement, documentation, pricing), verified Sep 25, 2026.
Where we tested it
Dragon
TaskThe model's task was to draw a picture of a dragon as an SVG.
Two dragons worth 29 and 24 points out of 50, a shared 14th and a 22nd place among 34. In describing its process it named four browsers, while the log shows only one.

The train test
TaskThe model had to play a scene with a courier on a night train and go through a fixed script of my ten turns.
8 and 7.5 to 8 points out of 8. In one run it let the stolen key open the cuff, even though its character had stopped the woman at the wall one turn earlier. Right after that it made the theft worthless with a lock of its own. The scoring key took nothing off for it, but only thanks to an exception that was created while scoring this very response.
You'll find the size band and figures for other models on the Tested models page.