Bolt Forge launches open models, 50x usage
Bolt.new launched Forge, an experimental agent running GLM 5.3, Kimi K3, and DeepSeek V4 Pro. Individual Pro users receive up to 50x more Forge usage through October 14 in exchange for optional anonymized prompts, code, and fix traces that help Arcee AI train an open-weight model. [Bolt announcement](https://bolt.new/blog/what-is-bolt-forge)
Forge is a smart usage-for-training-data exchange that could make open models materially better at real software development, but developers should treat it as an experimentation lane rather than a production default.
- –Bolt reports Forge scoring 92.2 versus 101.0 for its top paid model on the company’s own Build Index—a promising but self-reported gap.
- –The data trade is unusually explicit: consent is requested each time, but shared sessions include prompts, code, and debugging traces.
- –Switching back stops new sharing, though data already incorporated into training remains in the resulting models.
- –Dedicated hardware and browser-based WebContainers help explain how Bolt can offer dramatically higher usage at lower inference costs.
- –Bolt recommends duplicating projects first; Forge models remain experimental and currently lack PDF-upload support.
DISCOVERED
2h ago
2026-09-15
PUBLISHED
2h ago
2026-09-15
RELEVANCE
AUTHOR
SurKopu