A September 30 report summarized six post-launch Geekbench 7 runs associated with OpenAI Dots. Their reported median was about 8,550 multi-core points, compared with about 1,394 for Meta Muse. The configurations used nine and two CPU cores, respectively, so the roughly 6.1× score ratio is a comparison of benchmark results—not task-completion times.
Reported Geekbench medians for Dots and Meta Muse
Across the six Dots-associated runs, the reported scores ranged from 1,512 to 1,614 in single-core tests and from 8,135 to 8,991 in multi-core tests. The reported medians were 1,570 and about 8,550. Meta Muse’s medians came from 10 Geekbench records.
| Reported configuration | CPU cores | Geekbench 7 single-core median | Geekbench 7 multi-core median |
| OpenAI Dots-associated runs | 9 | 1,570 | About 8,550 |
| Meta Muse | 2 | About 1,041 | About 1,394 |
The multi-core median ratio is approximately 8,550 ÷ 1,394 ≈ 6.1. Geekbench CPU points measure performance in the benchmark’s workloads; task completion time is a different measure.
What the September 25 Geekbench record lists
A separate Geekbench result uploaded September 25, four days before OpenAI announced Dots, scored 1,667 in single-core and 9,435 in multi-core tests. It lists an AMD EPYC 9V74 processor, nine cores, a 2.60 GHz base frequency, 9.73 GB of memory and Ubuntu 24.04.3 LTS.
That September 25 result is distinct from the six post-launch runs summarized in the report, which were associated with Debian. The earlier score is higher than every multi-core score in the reported post-launch range.
What OpenAI says Dots are
OpenAI announced Dots on September 29 as persistent ChatGPT agents powered by GPT-6 Astra. The company says each Dot has its own cloud computer and can continue work between conversations using connected apps.
For US users, OpenAI’s Help Center describes texting as a limited beta for some Pro accounts, subject to account access.
