AlphaEvolve
In plain wordsA Google DeepMind system that combines AI language models with an evolution-like search to invent and improve computer programs, finding new results in mathematics and computing.
A Google DeepMindGoogle DeepMind, prop. n.: Google's main artificial intelligence laboratory, formed in 2023 by joining DeepMind and Google Brain. It builds the Gemini models and scientific AI systems such as AlphaFold. system for algorithm discovery that runs an evolutionary loop around language modelslarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code.: Gemini models (a fast one producing many candidate edits, a stronger one fewer but more considered ones) propose changes to a program, automated evaluators score every candidate, and the best variants are kept in a database that seeds the next round of mutations. Because it needs only a machine-checkable score, it suits problems in mathematics and computer science where quality can be measured automatically. Reported results include a procedure for multiplying 4×4 complex-valued matrices with 48 scalar multiplications, one fewer than Strassen's 1969 algorithm needs when applied recursively; better-than-known solutions to about a fifth of the more than 50 open problems it was tried on, including a new lower bound for the kissing number in 11 dimensions; a scheduling rule for Google's data centers that frees, on average, about 0.7% of the company's fleet-wide computing capacity; and faster kernels that shortened Gemini's training time by about 1%.
We gave AlphaEvolve the scheduler and a simulator as its evaluator, then let it run for a week.
Origin
Announced by Google DeepMind on 14 May 2025 in a blog post and white paper. It generalizes FunSearch (2023), which evolved single functions, to whole code files and multiple objectives. In November 2025 Bogdan Georgiev, Javier Gómez-Serrano, Terence Tao and Adam Zsolt Wagner reported testing it on 67 mathematical problems.
Usage note
Its discoveries are checked by the evaluator, not by the language model, which guards against hallucinationhallucination, n.: When an AI states something false or made up — a fake fact, quotation or source — fluently and confidently, as if it were true. but limits it to problems with reliable automatic scoring.
Related
- genetic programminggenetic programming, n.: Using simulated evolution to write computer programs: random programs compete at a task, and the best are combined and mutated over many generations until one does the task well.
- program synthesisprogram synthesis, n.: Getting a computer to write a program automatically from a description of what the program should do — a logical specification, a few input–output examples or a request in plain language.
- Darwin Gödel MachineDarwin Gödel Machine, prop. n.: An experimental AI coding assistant that improves itself by rewriting its own code, testing each new version and keeping a family tree of variants to build on.
- AI scientistAI scientist, n.: An AI system meant to do research largely on its own: reading the literature, coming up with ideas, running experiments and writing up the results.
- automated AI researchautomated AI research, n.: Using AI systems to do the work of AI researchers, such as running experiments and writing code for new models, so that AI helps build its own successors.
- open-endednessopen-endedness, n.: The ability of a system to keep producing new and interesting things without end — as evolution and human culture do — instead of stopping once it solves one fixed task.
Sources
- Google DeepMind (2025). AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms. 14 May 2025. deepmind.google/discover/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms
- Novikov, A. et al. (2025). AlphaEvolve: A coding agent for scientific and algorithmic discovery. Google DeepMind white paper.
Mentioned in
Google DeepMindGoogle DeepMind, prop. n.: Google's main artificial intelligence laboratory, formed in 2023 by joining DeepMind and Google Brain. It builds the Gemini models and scientific AI systems such as AlphaFold., AI scientistAI scientist, n.: An AI system meant to do research largely on its own: reading the literature, coming up with ideas, running experiments and writing up the results., evolutionary computationevolutionary computation, n.: Problem-solving by simulated evolution, much as breeders improve crops: keep many candidate answers, let the better ones produce slightly altered offspring, drop the worse, and repeat for many rounds., program synthesisprogram synthesis, n.: Getting a computer to write a program automatically from a description of what the program should do — a logical specification, a few input–output examples or a request in plain language., recursive self-improvementrecursive self-improvement, n.: An AI improving its own design or training, so that each improved version is better at making the next improvement. It is the engine usually imagined behind an intelligence explosion., AI co-scientistAI co-scientist, n.: An AI system built to work alongside human scientists, proposing and debating new research ideas and hypotheses that people then test in the lab., Darwin Gödel MachineDarwin Gödel Machine, prop. n.: An experimental AI coding assistant that improves itself by rewriting its own code, testing each new version and keeping a family tree of variants to build on., FunSearchFunSearch, prop. n.: A Google DeepMind method that discovered new mathematical constructions by having a language model repeatedly rewrite short computer programs, keeping the versions that scored best., genetic programminggenetic programming, n.: Using simulated evolution to write computer programs: random programs compete at a task, and the best are combined and mutated over many generations until one does the task well., open-endednessopen-endedness, n.: The ability of a system to keep producing new and interesting things without end — as evolution and human culture do — instead of stopping once it solves one fixed task., quality diversityquality diversity, n.: A family of evolution-inspired search methods that aim to collect many different good solutions — for example, many distinct ways for a robot to walk — rather than a single best one.