AlphaEvolve · Agentic Research Directory

AlphaEvolve

AlphaEvolve evolves programs through repeated proposals and executable evaluation. The paper reports a 48-scalar-multiplication construction for two 4×4 complex matrices, compared with 49 for recursive Strassen. The public repository contains mathematical results and verifier code, explicitly excluding the agent itself. These bounded constructions do not establish general scientific autonomy or independently replicated system performance.

L3Multi-step tool-using agentic workflow Source Preprint (not peer-reviewed) Outcome Author-reported

Evidence & verification

Source level

Preprint (not peer-reviewed)
Grounded in a preprint that had not completed peer review at the time the source was checked.

Outcome verification

Author-reported
Outcomes are reported by the team that built the system. The entry is source-checked against the team's own publications; not independently reproduced.

Sources checked

2026-10-19 (UTC)

Checked source version: 2506.13131v1. Claims refer to this manuscript snapshot; final journal/proceedings version parity was not verified. Author-reported; independent replication not established in this review.

Source-checked does not mean independently reproduced. See the Method page for how these labels are assigned.

Sources (2)

Research applicability

Documented mathematics research workflow, with source-version and evaluation limits retained.

Getting started

Read the primary manuscript and its limitations first. Check implementation access and terms separately; this review did not run an installation, API call to a research system, or experiment.

Maintenance

Source snapshot checked 2026-10-19. Current availability and experimental version parity are separate claims.

License

Research code/data/manuscript terms require separate verification; repository metadata alone is not a reuse clearance.

Workflow & human oversight

L3 applies only to the user-specified program-search workflow. Users provide initial code, evaluator code and optional configuration; authors inspect the resulting mathematical constructions. Automated scores constrain proposals; executable evaluation is required. The full agent implementation is not in the public result repository. Stops and safety are evaluator/configuration dependent; this record does not certify every possible evolved program or broader autonomous science.

Workflow level is not an evidence score; higher is not better evidence.

Approach

LLMs mutate programs; executable evaluators score candidates; a population database supplies feedback and parents for subsequent evolutionary proposals.

Outcomes (as reported by the originating team)

A reported construction multiplies 4×4 complex matrices using 48 scalar multiplications versus recursive Strassen’s 49. This is a bounded arithmetic construction, not a runtime ratio or a result over every field.

Limitations

Tasks require programmable evaluators. Public results/verifier notebook explicitly excludes the AlphaEvolve system code. Mathematical artifact verification is distinct from independent replication of agent performance; no code was executed here.

Reproducibility

Author-reported; independent replication not established in this review. Documentary review only; no candidate code, model, simulation or experiment was executed. A public repository or result artifact is not proof of replication.

Adoption requirements

Untested suggestion: inspect the pinned sources and validate data, evaluator assumptions, implementation availability and researcher controls before planning a bounded trial. No setup or trial was executed.

Comments

No comments yet.

Add a comment

Comments are held for moderation before appearing. Critical or corrective comments are welcome — especially evidence that an entry overstates what has been verified.