OpenRound is an AI-native coding assessment for hiring engineers. Plans start at $129/mo billed annually, with one free assessment included.
Researched and fact-checked by Bartu Cavusoglu · Updated October 26, 2026
OpenRound is a web-based coding assessment platform from Fabric, built for hiring what it calls AI-native engineers. Rather than banning AI from interviews, it lets candidates use AI coding agents and then reports on how they worked. The site positions it as a replacement for LeetCode-style screens, take-home projects and work trials.
The homepage argues that existing methods predate AI. It says LeetCode tests memorization, take-homes can be completed by pasting the brief into Claude, and work trials are expensive and do not scale. OpenRound responds with one session of about 90 minutes on a real engineering problem.
The site describes four steps. First, candidates work on an existing codebase matched to the employer's stack, with a ticket such as fixing a token refresh race condition. Second, they get complete access to AI coding agents in their CLI, similar to how they work on the job.
Third, the employer gets a play-by-play report on AI collaboration, taste and the completeness of the final solution. Fourth, interviewers hold a live discussion with the candidate inside OpenRound's web-based interview platform, covering trade-offs and thought process.
The evaluation report is the central feature. The sample shows an overall score with a recommendation, plus six dimensions: Analysis, Discovery, Planning, Judgement, Execution and AI Collaboration. It also lists completion time, such as 1h 12m, so interviewers can see how long the candidate took.
Other features include a library of 100+ problem statements, candidate emails and flexible scheduling, where candidates choose their own slot. Higher tiers add custom assessments, custom evaluations and custom AI models, while Enterprise adds ATS integration and custom branding.
OpenRound is a coding assessment platform from Fabric for hiring AI-native engineers. Candidates solve real engineering problems in a codebase using AI coding agents, and employers get a report on how they worked.
Starter is $129 a month billed annually (listed as $149 on monthly billing) for 5 assessments a month. Pro is $449 a month billed annually (listed as $499 monthly) for 20 assessments a month. Enterprise pricing is custom.
The pricing page says one free assessment is included, no credit card is required, and plans can be canceled anytime. Both plans offer a Start for Free button.
The homepage describes one session of 90 minutes. The pricing page lists assessment time as 60-90 minutes. The site says one session replaces two rounds of interviews.
Yes. Candidates have complete access to AI coding agents in an interactive CLI, which the site says is similar to how they work on the actual job. The report then shows how they collaborated with AI.
The sample report shows an overall score and a recommendation, plus scores for Analysis, Discovery, Planning, Judgement, Execution and AI Collaboration. It also gives a play-by-play of how the candidate worked with AI.
The Enterprise plan lists ATS integration. The sources do not name any supported ATS, so buyers should confirm compatibility with the sales team.
The pricing page lists the cost to candidates as $0. Employers pay through the subscription plans.