September 23, 2026
Bolt chooses speed for vibe coding: 80% of prompts went to two Flash models
Over one week, 80% of prompts in Bolt Forge went to two Flash models: GLM 5.3 Flash received 63%, while DeepSeek V4.1 Flash received 17% (Bolt, 23.09.2026). Bolt Lite costs $9 per month, while annual Individual Pro is listed at $25 per month with 10 million tokens monthly.
bolt.new
@boltdotnew
A week of open models in Bolt Forge. The picture is obvious. GLM 5.3 Flash: 63% DeepSeek V4.1 Flash: 17% Two Flash models. 80% of all prompts. The larger models were right there, on the list. The people building apps chose speed and kept choosing it. $9 per month: https://bolt.new/lite
· 1.1K views
On 18 August, Bolt offered only Claude internally, with no model choice. On 14 September, Forge added GLM 5.3 Flash and GLM 5.3. Kimi K3 and DeepSeek V4 Pro were experimental options.
A week later, the two Flash options captured 80% of prompts. Larger models remained on the list, but app builders chose speed more often.
Access and price. On Individual Pro, Forge is enabled through the agent picker with one confirmation. Lite requires a waitlist and a code from an email. Bolt promises up to 50× more AI usage compared with Standard mode when choosing Forge's most economical model.
In its own Bolt Build Index, Forge scored 92,2 against 101,0 for the best paid model. Bolt calls Forge experimental and keeps Standard or Max for complex production work.
Until 14 October 2026, Individual Pro includes Forge at no additional cost and retains the promised up to 50× usage.
