AI
3 min read

Replit Shifts Default AI Coding Provider to OpenAI, Citing Cost and Benchmark Metrics

Replit has made OpenAI, not Anthropic, the default model provider for Replit Agent. TechCrunch describes Replit as an AI coding platform that "has become a favorite of startup founders and developers." Speaking at TechCrunch Disrupt, Replit co-founder and CEO Amjad Masad said the decision was "based on both cost, partnership, and performance," TechCrunch reported. He said Replit uses an internal benchmark to assess metrics like performance, and also publishes a public benchmark. Update, Oct 1

Olivetti-style brochure spread with flow diagrams of processing paths and a software routing engine, illustrating Replit Agent's default model provider switch
Olivetti-style brochure spread with flow diagrams of processing paths and a software routing engine, illustrating Replit Agent's default model provider switch

Replit has made OpenAI, not Anthropic, the default model provider for Replit Agent. TechCrunch describes Replit as an AI coding platform that "has become a favorite of startup founders and developers."

Speaking at TechCrunch Disrupt, Replit co-founder and CEO Amjad Masad said the decision was "based on both cost, partnership, and performance," TechCrunch reported. He said Replit uses an internal benchmark to assess metrics like performance, and also publishes a public benchmark.

Update, Oct 19

TechCrunch has changed its headline. The article now runs as "Replit makes OpenAI its default over Anthropic, CEO says." An editor's note says it "was originally published on October 16 at 2:22 p.m. PT. It has been updated to include more context around Replit's partnerships with Anthropic and OpenAI. Its title has been updated, as well."

According to TechCrunch, after Disrupt Masad clarified that customers can still access Anthropic models on the platform, but that OpenAI's ChatGPT is now the default. "Replit remains model agnostic," Masad told TechCrunch over the weekend. "Our users can choose Anthropic's Claude or other supported models … OpenAI is now our preferred partner for powering Replit Agent, based on performance, economics, and the strength of our partnership."

The on-stage claim is disputed. TechCrunch reports that Masad's comments onstage "were rejected by insiders familiar with the partnership between Anthropic and Replit," and that one person familiar, who asked not to be named, said the partnership was still in place. TechCrunch also cites a late-September Replit blog post that names Claude Fable 5 and Claude Fable 5.1 as models running in Replit Agent's production environment. That post, Free the models: Harness design at the frontier, published Sep 29, 2026, includes a table titled "Delegation in Replit Agent production" that covers Fable 5, Fable 5.1 and GPT-6 Astra.

Replit's Model selector docs explain how to pick Claude. "Builders on Core and Pro can choose an available model in Power Mode or Max Mode." In the prompt box, select Agent Mode, select Power or Max, then open the Primary model menu and choose a model. The page lists Claude Sonnet 5 and Claude Sonnet 4.6 in Power Mode, and Claude Fable 5.1, Claude Opus 5.5, Claude Opus 4.8, Claude Opus 5 and Claude Opus 5 Fast in Max Mode. "Free Mode always uses intelligent model routing," and builders in Free Mode cannot choose a model manually. The docs add that "Model selection is rolling out, and the available models can vary by rollout, organization settings, and authorization," and that "The Fast model variants require a Pro or Enterprise plan."

To see how Replit compares with other app builders, read our guide to the best vibe coding tools. For coding agents that let you pick Claude or OpenAI models directly, see Cursor vs Claude Code vs Codex.

Inference Costs and Greenfield Code Generation

Before its deal with OpenAI, Replit relied on Anthropic's Claude models to run its platform, and Masad said onstage that Replit had been, until recently, one of Anthropic's top five customers, according to TechCrunch. "I think ChatGPT is better now," he said, and on the benchmarks he named two OpenAI models: "consistently, Sol and Astra perform better on building vibe-coded applications from scratch."

Cost was one of the three reasons Masad gave. TechCrunch notes that the switch could significantly reduce Replit's spend with Anthropic, since most customers stick with the default model rather than choosing one themselves.

In August 2026, Replit introduced Free Mode, powered by OpenAI's GPT-5.6 Luna, which OpenAI says lets anyone turn ideas into working software "without worrying about token costs." When a task needs more advanced reasoning, Replit can route it to GPT-5.6 Sol.

Enterprise Growth and Platform Flexibility

The vendor change coincides with a shift in Replit's revenue mix. Masad said enterprise was about 2% of revenue last year and about 15% this year, and that more than 50% of revenue is B2B, TechCrunch reported. Replit is not sharing ARR numbers yet.

OpenAI is now the default for Replit Agent, but Masad told TechCrunch that "Replit remains model agnostic" and that users can still choose Anthropic's Claude or other supported models.

Sources

AIOpenAIAnthropicReplitCoding Agents