AlphaEvolve is a Gemini-based evolutionary coding agent released by Google DeepMind. It combines the creative proposals of a large language model with verification by an automated evaluator inside an evolutionary framework to discover and optimize general-purpose algorithms. Rather than humans telling it the answer, it is given only the problem and the evaluation criteria, then generates candidate solutions, scores them itself, and evolves toward better ones.

How does AlphaEvolve work?

Gemini proposes code variations (candidate solutions), an automated evaluator actually runs them and scores them, and the higher-scoring candidates survive to become the starting point for the next generation. This propose → verify → select → repeat cycle runs thousands of times, reaching solutions that would be hard for a human to conceive. The core premise is that the problem must be automatically verifiable.

What did it achieve in matrix multiplication?

AlphaEvolve found an algorithm that computes the product of two 4×4 complex matrices using 48 scalar multiplications, surpassing for the first time Strassen (1969), which had been the best known for this setting. It was a record unbroken for 56 years, and because matrix multiplication underlies nearly all computation from AI training to graphics, the impact of the improvement is large.

What were its results in math and infrastructure?

Across more than 50 open problems in analysis, geometry, combinatorics, and number theory, about 75% matched the best known results and about 20% improved on the best known solutions. It also improved scheduling in Google's data centers to recover computing resources, and made a key kernel in Gemini's training 23% faster, cutting overall training time by 1%.

Is AlphaEvolve a case of recursive self-improvement (RSI)?

Because AlphaEvolve even optimized the training of the Gemini that underpins it, it comes close to a real-world case of recursive self-improvement (RSI), in which an AI improves the infrastructure of AI. However, its power comes from problems that can be automatically scored, so it does not transfer directly to domains where it is hard to verify a correct answer cleanly. Designing a verifiable evaluator is the true key.

AlphaEvolve Deep Dive: How an AI Broke a 56-Year-Old Math Record

Google DeepMind's AlphaEvolve is a concrete answer to the question of whether an AI can discover new algorithms on its own. Through an evolutionary loop in which Gemini proposes code and an automated evaluator verifies it, it overturned a matrix-multiplication record that had stood unbroken for 56 years since 1969. It also took on more than 50 open math problems and improved some of them, and it even optimized Google's own data centers and Gemini's training. This article lays out what AlphaEvolve is, how it works, what it achieved, and its connection to recursive self-improvement (RSI).

What Is AlphaEvolve?

AlphaEvolve is a "Gemini-based evolutionary coding agent" released by Google DeepMind. To discover and optimize general-purpose algorithms, it combines the creative proposals of a large language model with verification by an automated evaluator inside an evolutionary framework.

The key point is that "humans don't tell it the answer." Given only the problem and the evaluation criteria, AlphaEvolve generates candidate solutions, scores them itself, and evolves toward better ones.

How It Works

The way it works is evolution itself. Gemini proposes code variations (candidate solutions), an automated evaluator actually runs that code and scores it, and the higher-scoring candidates survive to become the starting point for the next generation.

As this cycle of "propose → verify → select → repeat" runs thousands of times, it reaches solutions that would be hard for a human to think of. The core premise is that the problem must be "automatically verifiable," which is why it's especially strong in algorithmic and mathematical domains where answers can be scored.

Breaking a 56-Year Matrix-Multiplication Record

Its most famous achievement is in matrix multiplication. AlphaEvolve found an algorithm that computes the product of two 4×4 complex matrices using 48 scalar multiplications, surpassing for the first time the method by Strassen (1969), which had been the best known for this setting.

This is a record that had stood unbroken for fully 56 years. Matrix multiplication underlies nearly all computation, from AI training to graphics, so a single improvement can ripple out into wide-ranging efficiency gains.

Results in Math and Infrastructure

In mathematics, too, AlphaEvolve took on more than 50 open problems across analysis, geometry, combinatorics, and number theory. In about 75% it found solutions matching the best known results, and in about 20% it improved on the best known solution.

Its real-world infrastructure results are also significant. It improved scheduling in Google's data centers to recover computing resources, and it made a key matrix-operation kernel in Gemini's training 23% faster, cutting overall training time by 1%. At Gemini's scale, 1% amounts to hundreds of thousands of GPU-hours.

The Connection to RSI — and the Limits

Notably, AlphaEvolve even optimized the training of the very Gemini that underpins it. That comes close to a real-world case of recursive self-improvement (RSI), in which an AI improves the infrastructure of AI.

That said, it's no cure-all. AlphaEvolve's power comes from "problems that can be automatically scored," so it doesn't transfer directly to domains where it's hard to verify a correct answer cleanly. In other words, designing a "verifiable evaluator" is the true key to this approach.

References: Google DeepMind — AlphaEvolve · VentureBeat