<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g" -->

---
title: Bolt is giving developers 50x more compute. But there’s...
description: Bolt.new launched Forge, a research preview through October 14 that gives Pro subscribers up to 50 times more usage of open weight coding models in exchange...
canonical: https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Bolt is giving developers 50x more compute. But there’s a catch. | daily.dev
og:description: Bolt.new launched Forge, a research preview through October 14 that gives Pro subscribers up to 50 times more usage of open weight coding models in exchange...
og:url: https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g
og:image: https://api.daily.dev/og/posts/BY59aFv8G.png
og:image:alt: Bolt is giving developers 50x more compute. But there’s a catch.
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Bolt is giving developers 50x more compute. But there’s a catch.

**[The New Stack](https://daily.dev/sources/newstack)** · 5 min read · 0 upvotes · 0 comments

## Summary

Bolt.new launched Forge, a research preview through October 14 that gives Pro subscribers up to 50 times more usage of open weight coding models in exchange for opting in to share anonymized coding sessions — prompts, source code, fix traces, and agent conversations. The data feeds a project with Arcee AI to train a trillion-parameter-class open-weight model, with the first training run starting in October and weights eventually released publicly. Forge runs open models (GLM 5.3 Flash, GLM 5.3, with Kimi K3 and DeepSeek v4 Pro as experimental options) on Bolt's own reserved infrastructure rather than proprietary APIs, scoring about 91% of Bolt's top paid model on its internal Bolt Build Index. Standard, Max, Teams, and Enterprise accounts are excluded; developers can stop future sharing but already-used training data remains embedded in the models.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://thenewstack.io/bolt-forge-training-data>

## Questions this post answers

### What is Bolt.new's Forge program and what do I have to give up to use it?

Forge is a research preview from Bolt.new offering Pro subscribers up to 50 times more usage of open weight coding models through October 14. In exchange, developers must opt in to sharing anonymized session data, including prompts, source code, fix traces, and agent conversations, which trains a trillion-parameter-class open-weight model built with Arcee AI. Standard, Max, Teams, and Enterprise accounts cannot participate.

_daily.dev keeps developers weighing AI coding tool trade-offs like Forge's usage-for-data exchange informed._

### Which models does Bolt's Forge use and how do they compare to Bolt's top paid model?

Forge currently runs GLM 5.3 Flash and GLM 5.3 as primary models, with Kimi K3 and DeepSeek v4 Pro available as experimental options. These open models scored 92.2 on Bolt's internal Bolt Build Index versus 101.0 for its top paid model, putting them at roughly 91% of that score, though this only measures performance inside Bolt's own benchmark.

_developers comparing open-weight and proprietary coding models can track benchmarks like this on daily.dev._

### If I leave Bolt's Forge program, does my previously shared coding data get removed from the trained model?

No, leaving Forge only stops new sessions from being shared going forward. Bolt states that any session data already used for training will remain embedded in the resulting models even after a developer opts out, since the anonymized data has already been incorporated into training runs with Arcee AI.

_developers deciding whether to share coding sessions for AI training can follow these data policy details on daily.dev._

## Similar posts on daily.dev

- [It's Swapping Season](https://daily.dev/posts/it-s-swapping-season-uhw8ifpqf) · Latent Space · 0 upvotes · 0 comments
- [The Enterprise Case for Open-Weight Models on Databricks](https://daily.dev/posts/the-enterprise-case-for-open-weight-models-on-databricks-swjf2cir1) · databricks · 1 upvotes · 0 comments

---

Tags: [#machine-learning](https://daily.dev/tags/machine-learning), [#open-source](https://daily.dev/tags/open-source), [#ai-coding](https://daily.dev/tags/ai-coding)

[View this post on daily.dev](https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"Bolt is giving developers 50x more compute. But there’s a catch.","url":"https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g"},"datePublished":"2026-09-15T18:56:12.663Z","dateModified":"2026-09-18T06:11:54.617Z","description":"Bolt.new launched Forge, a research preview through October 14 that gives Pro subscribers up to 50 times more usage of open weight coding models in exchange...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/52a1f355b1be8c7c6e9cd2a861959a92?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/52a1f355b1be8c7c6e9cd2a861959a92?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The New Stack","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The New Stack","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/newstack","url":"https://daily.dev/sources/newstack"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"machine-learning,open-source,ai-coding","timeRequired":"PT5M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The New Stack","item":"https://daily.dev/sources/newstack"},{"@type":"ListItem","position":3,"name":"Bolt is giving developers 50x more compute. But there’s a catch."}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/bolt-is-giving-developers-50x-more-compute-but-there-s-a-catch--by59afv8g#faq","mainEntity":[{"@type":"Question","name":"What is Bolt.new's Forge program and what do I have to give up to use it?","acceptedAnswer":{"@type":"Answer","text":"Forge is a research preview from Bolt.new offering Pro subscribers up to 50 times more usage of open weight coding models through October 14. In exchange, developers must opt in to sharing anonymized session data, including prompts, source code, fix traces, and agent conversations, which trains a trillion-parameter-class open-weight model built with Arcee AI. Standard, Max, Teams, and Enterprise accounts cannot participate. daily.dev keeps developers weighing AI coding tool trade-offs like Forge's usage-for-data exchange informed."}},{"@type":"Question","name":"Which models does Bolt's Forge use and how do they compare to Bolt's top paid model?","acceptedAnswer":{"@type":"Answer","text":"Forge currently runs GLM 5.3 Flash and GLM 5.3 as primary models, with Kimi K3 and DeepSeek v4 Pro available as experimental options. These open models scored 92.2 on Bolt's internal Bolt Build Index versus 101.0 for its top paid model, putting them at roughly 91% of that score, though this only measures performance inside Bolt's own benchmark. developers comparing open-weight and proprietary coding models can track benchmarks like this on daily.dev."}},{"@type":"Question","name":"If I leave Bolt's Forge program, does my previously shared coding data get removed from the trained model?","acceptedAnswer":{"@type":"Answer","text":"No, leaving Forge only stops new sessions from being shared going forward. Bolt states that any session data already used for training will remain embedded in the resulting models even after a developer opts out, since the anonymized data has already been incorporated into training runs with Arcee AI. developers deciding whether to share coding sessions for AI training can follow these data policy details on daily.dev."}}]}
```

