Does Agents Last Exam work in the EU?

Unverified. No explicit evidence found. The terms say prize winners outside the US are responsible for the legality of accepting prizes in their country, and the privacy page says data may be processed outside the user's country. Last verified 14 Nov 2026.

How much does Agents Last Exam cost?

Pricing unknown. No pricing information found; participation is through a contributor program.

Where was Agents Last Exam shown?

In 1 creator video by AI Search. Most recently by AI Search on 14 Jun 2026, at 12:23 in "RIP Claude Fable, open-source AI unleashed, full body avatars, new Google models, new TTS: AI NEWS".

Agents Last Exam

Website

Benchmark of AI agents on long-horizon professional workflows across 55 sub-industries, led by Berkeley RDI.

UnverifiedResearch Agents & AutomationResearch

Updated 14 Nov 2026EU status checked 14 Nov 2026Last shown 14 Jun 2026

Agents Last Exam at a glance

What is Agents Last Exam?: Agents Last Exam is a research. Benchmark of AI agents on long-horizon professional workflows across 55 sub-industries, led by Berkeley RDI.
Does Agents Last Exam work in the EU?: Unverified. No explicit evidence found. The terms say prize winners outside the US are responsible for the legality of accepting prizes in their country, and the privacy page says data may be processed outside the user's country. Last verified 14 Nov 2026.
How much does Agents Last Exam cost?: Pricing unknown. No pricing information found; participation is through a contributor program.
Where was Agents Last Exam shown?: In 1 creator video by AI Search. Most recently by AI Search on 14 Jun 2026, at 12:23 in “RIP Claude Fable, open-source AI unleashed, full body avatars, new Google models, new TTS: AI NEWS”.

Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.

About

Agents' Last Exam is a benchmark that evaluates AI agents on long-horizon, economically valuable professional tasks with verifiable outcomes. It is led by Berkeley RDI together with 300+ industry experts, and its site states that it covers all 55 targeted sub-industries of professional computer-based work. Over 1,500 tasks have been collected toward a 5,000-task target.

Tasks are run in real professional software, for example Adobe After Effects (motion and VFX), Siemens NX (3D modeling), Unreal Engine (game development), Moldex3D (mold flow analysis), Rhino 3D (architectural modeling and energy analysis) and FSLeyes (brain imaging). The stated aim is objective, comparable scores across domains.

The project also runs a contributor program. Domain experts can submit workflow data without coding, and researchers and engineers can turn workflows into reproducible benchmarks. Qualifying contributors are offered co-authorship, and high-impact contributions can receive awards from a $100K+ funding pool. A YouTube review reported GPT-5.5 with Codex at the top of its results, ahead of Claude Fable 5. This has not been verified against the official pages. The scraped pages do not show a public paper, code repository or dataset release.

Key features

1,500+ tasks collected toward a 5,000-task target
Covers 55 sub-industries of professional computer-based work
Tasks run in real professional software such as After Effects, Siemens NX, Unreal Engine and Rhino 3D
Verifiable outcomes for objective, comparable scoring
Led by Berkeley RDI with 300+ industry experts
Contributor program with co-authorship and a $100K+ award pool

1 moment

As seen on YouTube

Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.

AI Search14 Jun 2026

RIP Claude Fable, open-source AI unleashed, full body avatars, new Google models, new TTS: AI NEWS

Benchmark of long professional workflows across 55 sub-industries. GPT-5.5 with Codex ranks top, beating Claude Fable 5, followed by GPT-5.5 with OpenClaw and Cursor's Composer.

12:23Open on YouTube