Bolt.new stumbles against Lovable, Replit

// 1h agoBENCHMARK RESULT

Bolt.new stumbles against Lovable, Replit

ANNOUNCEMENT PRODUCT PRODUCT HUNT YOUTUBE

A head-to-head vibe-coding test finds Bolt.new unable to finish a project-tracking app that Lovable and Replit complete successfully. That result cuts against Bolt’s pitch as a browser-based, prompt-to-deployed full-stack builder with built-in databases, authentication, and hosting. [Bolt.new](https://bolt.new/use-cases/ai-software-development)

// ANALYSIS

Bolt.new remains compelling for rapid scaffolding, but this is a meaningful reliability miss: a polished preview means little if core workflows fail under a realistic prompt.

–Project tracking stresses CRUD operations, navigation, relationships, persistence, and multi-screen state—the areas where generated full-stack apps often break.
–Lovable and Replit completing the same task makes this a comparative capability signal, not merely an isolated Bolt bug.
–Bolt’s browser-native workflow, built-in infrastructure, and WebContainers remain genuine advantages for fast prototypes. [Bolt documentation](https://support.bolt.new/get-started/intro-bolt)
–Developers should treat Bolt output as a starting point requiring flow-by-flow testing, especially before trusting it with production data or team workflows.

// TAGS

bolt-newai-codingcoding-agentcode-generationno-codebenchmarkdevtool

DISCOVERED

1h ago

2026-10-09

PUBLISHED

1h ago

2026-10-09

RELEVANCE

8/ 10

AUTHOR

Rob The AI Guy

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

MODEL35m ago

Microsoft-Decision-1 brings fast decisions to Foundry

Microsoft released Microsoft-Decision-1, a model for fast, structured decision-making rather than open-ended text generation. Microsoft says it is available in Foundry and is being tested for incident response, quality control, and scientific discovery.

LAUNCH58m ago

Matter Compiler Turns Code Into Micro-Machines

Atomic Machines emerged from six years in stealth with $250 million raised to date and introduced the Matter Compiler, an AI-native manufacturing system that builds functional micro-machines from digital instructions. Its first product, PrimeSwitch, is a fast, low-resistance power relay designed for AI data-center infrastructure.

BENCHMARK1h ago

Microsoft CABRA Exposes Coding-Agent Blind Spots

Microsoft’s CABRA is an open-source benchmark that generates synthetic code-editing tasks from call graphs, scaling difficulty across traversal, search, runtime reasoning, and instruction following. Its 6,840-task evaluation finds that tools preserve agent performance until tasks demand deeper code understanding.