Alpha and Beta testing are forms of

Concept

Software-testing terms sit on two independent axes. The design-technique axis describes how test cases are derived — from the program's internal structure, or from its specification alone. The level axis describes the stage of the lifecycle — unit, integration, system, then acceptance — fixing who tests, against which standard, and at what point.

Applying it here

1. Alpha and beta testing are defined by when they happen and who performs them, not by how their test cases are derived, so they belong on the level axis.

2. Alpha testing is run at the developer's site by internal users or a small invited group, in a controlled environment, before the software is released.

3. Beta testing is run at the users' own sites by selected external users under real operating conditions, and their feedback is collected before general release.

4. In both, the standard being judged is the users' own requirements and their readiness to take delivery — and that is precisely the acceptance level.

Hence Alpha and Beta testing are forms of Acceptance Testing.

Cross-check against the other categories

Alpha and beta tests are commonly run using black-box techniques, and beta testing in particular is carried out by external users who have no access to the source code. Technique and level are separate axes, however: whichever technique is applied, the classification this question asks for is the level, and that level is acceptance testing.