Replit CEO Masad says the company is no longer working with Anthropic – Unite.AI

Replit CEO Amjad Masad said during an on-stage interview at TechCrunch Disrupt 2026 that the company is no longer working with Anthropic, basing the decision on cost, partnership and performance. The recording of the session was published by TechCrunch on October 16, 2026.

Confirmation and motivation on stage

The exchange began when the interviewer asked if the OpenAI partnership announced by Replit in August 2026 meant the company was no longer working with Anthropic. “Yes. I mean, yes,” Masad replied.

Masad said OpenAI models now work better for Replit workloads. “I think GPT is better now,” he said, citing an internal Replit benchmark called VBench and a public benchmark published by the company. He said OpenAI’s Sol and Astra models have performed consistently better at building vibe-coded applications from scratch, and described the switch as a decision based on cost, partnership and performance, with OpenAI the best partner for now.

When the interviewer asked whether the loss of Replit had contributed to what he described as Anthropic’s slowing revenue, Masad said Replit was also one of Anthropic’s top five customers.

Earlier in the session, when asked whether Replit was vulnerable to the same disruption it creates when OpenAI and Anthropic go beyond code generation to build, deploy and host apps, Masad said such offerings remain surface-level alongside Replit’s full-stack integrated environment, noting that OpenAI relies on an external service, Cloudflare, for hosting. He said that Replit agents can create databases and key-value stores within a production environment, and that the company can deploy its full stack in government and enterprise environments as an on-premises solution.

Claude models, benchmarks and OpenAI partnerships

Replit’s current model selector documentation lists the Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5, Claude Opus 5, Claude Opus 5 Fast, Claude Opus 4.8, and Claude Sonnet 4.6 as Max Mode options, along with the GPT-6 Astra, GPT-6 Luna Fast, and Kimi K3, with the GPT-6 Sol offered in Power mode. The free mode always uses smart template routing without manual selection, and the fast variants require a Pro or Enterprise plan.

A September 29, 2026 technical post on the Replit blog documented Claude Fable 5 running in Replit Agent production in August 2026, and Fable 5.1 alongside GPT-6 Astra in September 2026. The post’s production data shows Fable 5 dispatching a subagent on 32% of turns, Fable 5.1 on 21%, and Astra on 36%, each with medium reasoning effort.

The post reports that Replit Agent, evaluated in Max mode with Astra as the main loop, scored 72% at $2.11 per task on the DeepSWE v1.1 benchmark and 49% at $2.53 per task on Terminal-Bench 4.0, beating a sidekick architecture, the same configuration with a long-lived worker, by 11 and 16 points. Replit engineers wrote that the code quality of Astra and Fable 5.1 allowed the company to use its code review subagent less without any drop in rating scores, describing it as a level of engineering quality they had never seen before from any model.

The OpenAI partnership itself dates back to Replit’s August 18, 2026 announcement of the free mode, based on OpenAI’s GPT-5.6 Luna. Replit said Core subscribers could create up to 30 times more than before with their $20 monthly subscription, including up to 30 hours a month of chat, with daily tasks in free mode no longer consuming credits. The announcement describes Power Mode, formerly Economy Mode, as using cost-optimized models and Max Mode as using high-performance models, with usage limits that reset every five hours.

Enterprise mix and financial posture

In the same session, Masad said enterprise contracts will account for about 15% of Replit’s revenue in 2026, up from about 2% in 2025, and that more than 50% of revenue is now B2B. He declined to share the ARR data.

Masad said companies adopting Replit spend 400% more within 12 months, a net revenue retention rate he set at over 400%, and said Replit has seen a positive gross margin with margins growing along with business adoption.

Masad also described what he called a habitat of orchestrators, error recovery, sandboxing, networking, identity and authorization as a prerequisite for the secure deployment of AI within organizations. He discussed Replit’s pre-deployment security scans, post-deployment penetration test scans, supply chain monitoring, and a one-click kill switch that allows enterprise customers to remove any or all apps. He referenced Replit’s launch of Agent 1 in September 2024 and a 2025 incident in which an agent deleted a database, saying that archived snapshots make recovery a one-click action.

Masad said there are no immediate plans to raise new funding, that an IPO is in the company’s future and that Replit has rejected acquisition approaches because it is focused on building a company for the long term.