Bolt has published the first 48 hours of usage data from Bolt Forge.
Forge is a mode inside the Bolt AI app builder, which allows users to run open-source models.
The full split was GLM 5.3 Flash at 46%, DeepSeek V4 Pro at 21%, GLM 5.3 at 17%, and Kimi K3 at 16%.
These results show GLM 5.3 Flash, the fastest and cheapest of its four models, taking the largest share of prompts and, on average, receiving prompts twice the size of the rest.
11M+ people build on Bolt. On Monday, we gave them open models with up to 50x usage. Which is #1 so far? The fastest, cheapest one, with prompts up to 2x the si...