Wagtail reviews a month of coding with GLM 5.3 Flash
What happened
On October 2, Hacker News featured the Wagtail team's review of a month spent coding with GLM 5.3 Flash. The team tried to use the model for a full month of coding work, but the target model ended up at only 50% of usage: 2 billion tokens total, 1 billion of them from GLM 5.3 Flash. The report also discloses energy use of about 35 kWh, above the 10 kWh target. The review gives the model's actual share of usage and the energy result, showing the month-long trial did not put all usage on the target model and missed the energy target.
Written by AI from the coverage · updated 1 hour ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageOne month coding with GLM 5.3 Flash
Wagtail 团队尝试整月使用 GLM 5.3 Flash 编码，最终目标模型仅占 50% 用量，即 2B tokens 中的 1B，能耗约 35 kWh，高于 10 kWh 的目标。
Heat over time
Not enough continuous observations to draw a trend yet.