Benchmarking AI with Stanislav Khromov

About this episode

Stanislav Khromov discusses Svelte Bench and the challenges of AI-assisted coding with Svelte 5.

In this episode, Stanislav Khromov joins the Svelte Radio team to discuss his work on Svelte Bench, a benchmarking tool that scientifically measures how well different LLMs understand and write Svelte 5 code. The conversation explores the challenges of AI-assisted coding with Svelte 5, the development of the official Svelte MCP (Model Context Protocol) that provides LLMs with documentation and auto-fixing capabilities, and the surprising performance differences between major AI providers. Stanislav also shares insights on working with AI tools, the future of local models, and the economics of AI coding assistants. Notes Stanislav Khromov Svelte Bench GitHub Svelte MCP Anthropic Spending VC Money Unpopular Opinions Kevin: Serverless is overrated Stanislav: OpenAI has the worst models of the big three Antony: Heat Pumps are bad Technology Connections Tech Ingredients Picks Kevin: Kasai Hibachi Grill (now tested, and can confirm, it is PENG!) Stanislav: Ball x Pit Antony: Saunas By the Sea

People in this episode

Host: Kevin Åberg Kultalahti

Guest: Stanislav Khromov

Topics covered

• AI
• Svelte 5
• benchmarking
• coding assistants
• local models
• performance differences

Keywords

• Svelte
• AI
• benchmarking
• LLMs
• coding assistants
• MCP
• performance

Mentioned in this episode

Organizations: Svelte Bench, Svelte MCP, OpenAI, Anthropic

Products: Kasai Hibachi Grill, Ball x Pit, Saunas By the Sea

More episodes of Svelte Radio

• npmx shows what npmjs won't · March 6, 2026 · 57 min
• How did Svelte do in the State of JavaScript 2025?! · February 10, 2026 · 53 min
• The LLM Doomer Episode · February 4, 2026 · 41 min
• Custom Renderers for Svelte with Paolo Ricciuti · December 11, 2025 · 1h 1m
• One Man Svelte for NGOs with Ilja · December 6, 2025 · 1h 14m
• Macrodata Refinement with Ken Kunz · October 23, 2025 · 1h 5m