Compare· two models, one test
Compare two models
| Anthropic: Claude Opus 5.5 S | OpenAI: GPT-6.1 Sol S | |
|---|---|---|
| MRMP score | 98.59 | 94.42 |
| Report card | 39 / 40 | 38 / 40 |
| Showcase | 59.59 / 60 | 56.42 / 60 |
| Mohsen’s take | – | – |
| Cost of answers | $5.12 | $1.01 |
| Speed | 137 tokens/s | 59 tokens/s |
| Runs | 2 | 2 |
| Simorgh Flight | 7.5 | 7.5 |
| Ant Colony | 7.1 | 5.4 |
| 3D Azadi Tower | 7.5 | 7.5 |
| Virtual Santur | 7.5 | 6.4 |
| Autumn Book Café | 7.5 | 7.1 |
| Persian Calendar | 7.5 | 7.5 |
| Girih Tiles | 7.5 | 7.5 |
| Bazaarcheh | 7.5 | 7.5 |
Anthropic: Claude Opus 5.5 · 7.50 of 7.5
OpenAI: GPT-6.1 Sol · 7.50 of 7.5
Scores are each model’s mean over its runs; the builds are from its newest run. Costs are in US dollars.