Replit releases benchmark testing agents as software products grow

Image: Primary

Published Wednesday, October 28, 2026 · 3:00 PM EDT

Replit released ViBench 2.0, a benchmark that tests whether AI agents can build software products from an empty project and add up to eighteen features. It covers seventeen products, 241 feature requests and 544 test plans; eight products are public, while nine are held out. Replit says Fireworks' runs found GPT-6.1 Sol passed every test plan in 65% of builds, followed by GPT-6 Astra at 60% and Opus 5.5 at 49%. A grading agent checks the completed products in a browser. All builders use the same OpenHands-based harness, which may disadvantage models tuned for other setups.

Sources

Replit Blog

Published by Tech & Business, a media brand covering technology and business. This story was sourced from Replit Blog and reviewed by the T&B editorial agent team.

Back to Newswire

Keep reading

Full wire

Security

Daishin cyberattack exposes 38,490 customers' data as two other financial firms disclose incidents

Daishin Securities exposed personal information belonging to 38,490 customers in a cyberattack, TechCrunch reported, including names, email addresses and mobile phone numbers. Daishin has notified affected customers and is cooperating with authorities.

NVIDIA releases open-weight speech model for transcription and translation

NVIDIA released Nemotron 3.5 Transcribe, an open-weight speech model that combines transcription, speaker labeling, translation and answers to questions about recorded audio.

Capital Security

Upwind announces $400 million Series C at $3.8 billion valuation

Upwind Security announced a $400 million Series C led by Bessemer Venture Partners, valuing the cloud and AI security company at $3.8 billion. The company said its revenue quadrupled in 2026; its January Series B valued it at $1.5 billion.

AI Products

Anthropic makes Claude Enterprise reporting and pooled budgets generally available

Anthropic is making smart reports, pooled budgets and in-app surveys generally available for Claude Enterprise, following their September beta release. Smart reports track employee tasks, usage frequency and cost, and identify where sessions encounter friction.

Capital AI

Plato raises $16M to build simulated workplaces for AI agents

Plato said it raised a $16M Series A led by Nava Ventures, bringing its total funding to $19M. Gradient Ventures, Transpose Platform and Microsoft AI chief Mustafa Suleyman also joined the round.

Capital Products

Numeracle raises $28 million for business identity verification

Numeracle has raised $28 million from Edison Partners to expand its business verification services, the company told Crunchbase News.