
AlphaEvolve
Gemini-powered evolutionary coding agent that discovers algorithms better than human designs.
About
Unveiled May 14, 2025, AlphaEvolve is Google DeepMind's general-purpose coding agent for algorithm discovery. It pairs Gemini models (Flash for breadth, Pro for depth) with an evolutionary loop: the LLM proposes code mutations, automated evaluators score them, and the best candidates seed further generations. Its headline result was a new 4x4 complex matrix-multiplication algorithm using 48 scalar multiplications, beating a record that had stood since Strassen's 1969 work.
AlphaEvolve has been deployed inside Google itself, reportedly recovering 0.7% of global data-center compute through better scheduling heuristics and speeding up Gemini training kernels, and it is now generally available on Google Cloud. The honest caveats: it is a research/enterprise system with no public weights, and discovered algorithms still need human verification and engineering before production use — a published construction is not the same as a deployed optimization.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 70; Trending score 85/100 from researcher assessment.