Hermes Agent used 10M tokens for weak basic automation
was installed on an old MacBook and connected to so several models could be tested. In about two days, it used roughly 10 million tokens, but the output quality was still disappointing. Telegram was connected, user-md was configured through an interview flow, and two or three cronjobs were set up through the .
The jobs were simple news-fetching and summary tasks on different topics. Smaller models handled basic chat somewhat well, but fell apart when cronjobs needed . Qwen3.6:35B, DS V4 Flash, and Super 122B gave uneven results.
A larger model such as GLM5.2 produced more acceptable output, but token use became very high for such basic work. Model choices and other settings also did not appear to stay saved reliably within or across sessions.
Key points
- About 10 million tokens were used in roughly two days for mediocre results.
- ran on an old MacBook with connected for model testing.
- Telegram, user-md, and cronjobs were configured through the .
- Smaller models struggled when cronjobs needed , even if basic chat was acceptable.
- GLM5.2 gave better output, but token use was very high for basic summaries.