| 2026-07-22 |
[Counterexample to the Dinitz–Garg–Goemans conjecture](/finding/2026-07-22-dinitz-garg-goemans) |
Independent | GPT-5.6 Pro | Claimed | AI-led | computer-science |
| 2026-07-20 |
[Counterexamples to the Gaussian moments conjecture](/finding/2026-07-20-gaussian-moments) |
Independent | GPT-5.6 Sol Pro + Claude Fable 5 | Author verified | AI-led | mathematics |
| 2026-07-19 |
[Counterexample to the Jacobian conjecture in dimension three](/finding/2026-07-19-jacobian-conjecture) |
Anthropic | Claude Fable 5 | Formally verified | Collaborative | mathematics |
| 2026-07-14 |
[Near-quadratic lower bound for derivative-free convex optimization](/finding/2026-07-14-zeroth-order-oracle-bound) |
Independent | GPT-5.6 Sol Pro | Formally verified | AI-led | mathematics |
| 2026-07-11 |
[Counterexample to Grothendieck's question on finite flat group schemes](/finding/2026-07-11-grothendieck-group-schemes) |
Independent | OpenAI Sol (construction); Claude Fable (Lean formalisation) | Formally verified | Collaborative | mathematics |
| 2026-06-25 |
[First Herculaneum scroll read end to end without unrolling it](/finding/2026-06-25-herculaneum-scroll-read) |
Vesuvius Challenge | Community-developed ink-detection neural networks | Author verified | AI-assisted | archaeology |
| 2026-05-21 |
[Nine Erdős problems and 44 OEIS conjectures proved with machine-checked proofs](/finding/2026-05-21-alphaproof-nexus) |
Google DeepMind | AlphaProof Nexus (LLM + Lean) | Formally verified | Search scaffold | mathematics |
| 2026-05-20 |
[Disproof of the Erdős unit-distance conjecture](/finding/2026-05-erdos-unit-distance) |
OpenAI | GPT-5 series reasoning model | Independently checked | AI-led | mathematics |
| 2026-01-15 |
[Structure in Bruhat intervals of permutation groups](/finding/2026-01-alphaevolve-bruhat) |
Google DeepMind | AlphaEvolve (Gemini-based) | Independently checked | Search scaffold | mathematics |
| 2026-01-13 |
[Erdős problem #728 resolved and formalized in Lean](/finding/2026-01-erdos-728) |
OpenAI / Harmonic | GPT-5.2 Pro + Aristotle | Formally verified | Autonomous | mathematics |
| 2025-11-20 |
[Early science acceleration experiments with GPT-5](/finding/2025-11-gpt5-science-acceleration) |
OpenAI | GPT-5 | Peer reviewed | AI-assisted | computer-science |
| 2025-11-03 |
[AlphaEvolve across 67 problems: 20 improvements, 8 regressions](/finding/2025-11-03-alphaevolve-at-scale) |
Google DeepMind | AlphaEvolve (Gemini-based) | Independently checked | Search scaffold | mathematics |
| 2025-10-19 |
[GPT-5 "solved 10 Erdős problems": it located existing solutions](/finding/2025-10-19-gpt5-erdos-retrieval) |
OpenAI | GPT-5 | Already known | Retrieval | mathematics |
| 2025-09-25 |
[Limits to black-box amplification in QMA, with the key step written by GPT-5](/finding/2025-09-25-qma-amplification-limits) |
UT Austin / CWI Amsterdam | GPT-5 Thinking | Author verified | AI-assisted | computer-science |
| 2025-09-17 |
[New families of unstable singularities in fluid equations](/finding/2025-09-17-unstable-singularities) |
Google DeepMind (with Brown, NYU and Stanford) | Physics-informed neural networks with Gauss–Newton optimisation | Author verified | Search scaffold | mathematics |
| 2025-09-17 |
[AI-generated bacteriophage genomes that replicate and kill bacteria](/finding/2025-09-17-evo-phage-genomes) |
Arc Institute / Stanford University | Evo 1 and Evo 2 | Author verified | AI-led | biology |
| 2025-09-11 |
[Strong prime number theorem formalized in Lean by an autoformalization agent](/finding/2025-09-11-gauss-strong-pnt) |
Math, Inc. | Gauss (autoformalization agent) | Author verified | AI-assisted | mathematics |
| 2025-08-01 |
[Improved step-size bound in smooth convex optimization](/finding/2025-08-gpt5-convex-bound) |
OpenAI | GPT-5 Pro | Already known | AI-assisted | mathematics |
| 2025-07-28 |
[Machine-checked Lean proofs for five of six 2025 IMO problems](/finding/2025-07-28-aristotle-imo-lean) |
Harmonic | Aristotle | Formally verified | AI-led | mathematics |
| 2025-07-21 |
[Gold-medal standard at the 2025 International Mathematical Olympiad](/finding/2025-07-21-gemini-deepthink-imo) |
Google DeepMind | Gemini Deep Think (advanced version) | Independently checked | AI-led | mathematics |
| 2025-06-03 |
[Phase 2a results for a drug whose target and molecule both came from AI](/finding/2025-06-03-rentosertib-phase2a) |
Insilico Medicine | PandaOmics (target) + Chemistry42 (molecule) | Peer reviewed | Search scaffold | medicine |
| 2025-05-14 |
[4×4 complex matrix multiplication in 48 scalar multiplications](/finding/2025-05-alphaevolve-matmul) |
Google DeepMind | AlphaEvolve (Gemini-based) | Independently checked | Search scaffold | computer-science |
| 2025-05-14 |
[Improved lower bound for the 11-dimensional kissing number](/finding/2025-05-alphaevolve-kissing) |
Google DeepMind | AlphaEvolve (Gemini-based) | Independently checked | Search scaffold | mathematics |
| 2025-05-14 |
[Improved bound for the Erdős minimum-overlap problem](/finding/2025-05-alphaevolve-minimum-overlap) |
Google DeepMind | AlphaEvolve (Gemini-based) | Independently checked | Search scaffold | mathematics |
| 2025-05-01 |
[Candidate treatment for dry age-related macular degeneration](/finding/2025-robin-macular) |
FutureHouse | Robin (multi-agent) | Claimed | AI-led | biology |
| 2025-03-12 |
[A fully machine-generated paper passed workshop peer review](/finding/2025-03-12-ai-scientist-workshop-paper) |
Sakana AI | The AI Scientist-v2 | Author verified | AI-led | computer-science |
| 2025-02-19 |
[AI co-scientist hypotheses on antimicrobial resistance and liver fibrosis](/finding/2025-02-ai-coscientist-amr) |
Google | AI Co-Scientist (Gemini 2.0 multi-agent) | Author verified | AI-assisted | biology |
| 2025-02-13 |
[Enzymes with working catalytic machinery designed from scratch](/finding/2025-02-13-designed-serine-hydrolases) |
Institute for Protein Design, University of Washington | RFdiffusion with PLACER ensemble scoring | Peer reviewed | Search scaffold | chemistry |
| 2025-01-16 |
[A working fluorescent protein generated by a language model](/finding/2025-01-16-esmgfp) |
EvolutionaryScale | ESM3 (98B) | Peer reviewed | AI-led | biology |
| 2025-01-16 |
[Generative model designs crystals to order; its flagship synthesis turned out to be a known compound](/finding/2025-01-16-mattergen) |
Microsoft Research | MatterGen | Disputed | Search scaffold | materials |
| 2024-11-20 |
[A neural decoder that identifies quantum errors more accurately than hand-designed methods](/finding/2024-11-20-alphaqubit) |
Google DeepMind / Google Quantum AI | AlphaQubit | Peer reviewed | AI-led | physics |
| 2024-10-02 |
[Complete wiring diagram of an adult fruit-fly brain](/finding/2024-10-02-flywire-connectome) |
FlyWire Consortium (Princeton, MRC LMB, Cambridge, Vermont) | Automated electron-microscopy segmentation networks | Peer reviewed | AI-assisted | neuroscience |
| 2024-07-25 |
[Silver-medal standard at the 2024 International Mathematical Olympiad](/finding/2024-07-25-alphaproof-imo) |
Google DeepMind | AlphaProof + AlphaGeometry 2 | Independently checked | AI-led | mathematics |
| 2024-05-08 |
[Joint structure prediction for proteins, nucleic acids and ligands (AlphaFold 3)](/finding/2024-05-08-alphafold3) |
Google DeepMind / Isomorphic Labs | AlphaFold 3 | Peer reviewed | AI-led | biology |
| 2024-02-21 |
[Reinforcement learning steers a tokamak away from tearing instabilities](/finding/2024-02-21-tearing-mode-avoidance) |
Princeton University / PPPL / DIII-D National Fusion Facility | Deep reinforcement learning controller over a learned plasma model | Peer reviewed | Search scaffold | physics |
| 2024-01-17 |
[Olympiad geometry solved without human demonstrations](/finding/2024-01-17-alphageometry) |
Google DeepMind | AlphaGeometry | Peer reviewed | Search scaffold | mathematics |
| 2023-12-20 |
[A language-model agent that planned and ran chemistry experiments on lab robots](/finding/2023-12-20-coscientist) |
Carnegie Mellon University | GPT-4 (Coscientist agent) | Peer reviewed | AI-led | chemistry |
| 2023-12-20 |
[A new structural class of antibiotic candidates against MRSA](/finding/2023-12-20-antibiotic-structural-class) |
MIT / Broad Institute / Harvard | Ensembles of graph neural networks with substructure attribution | Peer reviewed | Search scaffold | chemistry |
| 2023-12-14 |
[New lower bound constructions for the cap set problem](/finding/2023-12-funsearch-capset) |
Google DeepMind | FunSearch (PaLM 2 / Codey) | Peer reviewed | Search scaffold | mathematics |
| 2023-12-14 |
[Improved heuristics for online bin packing](/finding/2023-12-funsearch-binpacking) |
Google DeepMind | FunSearch (PaLM 2 / Codey) | Peer reviewed | Search scaffold | computer-science |
| 2023-11-29 |
[Large-scale prediction of new stable crystalline materials (GNoME)](/finding/2023-11-29-gnome) |
Google DeepMind | GNoME (graph neural network) | Disputed | Search scaffold | materials |
| 2023-11-29 |
[Autonomous laboratory reports solid-state synthesis of new inorganic compounds](/finding/2023-11-29-a-lab-synthesis) |
Lawrence Berkeley National Laboratory | A-Lab (ML planning + robotics) | Disputed | AI-led | materials |
| 2023-11-14 |
[Medium-range weather forecasts from a graph neural network beat the operational physics model](/finding/2023-11-14-graphcast) |
Google DeepMind | GraphCast | Peer reviewed | AI-led | climate |
| 2023-09-19 |
[Pathogenicity predictions for 71 million human missense variants](/finding/2023-09-19-alphamissense) |
Google DeepMind | AlphaMissense | Peer reviewed | AI-led | medicine |
| 2023-06-07 |
[Faster sorting routines discovered and merged into the LLVM C++ library](/finding/2023-06-07-alphadev) |
Google DeepMind | AlphaDev (AlphaZero-based) | Independently checked | Search scaffold | computer-science |
| 2022-10-05 |
[Faster matrix-multiplication algorithms found by reinforcement learning](/finding/2022-10-05-alphatensor) |
Google DeepMind | AlphaTensor (AlphaZero-based) | Peer reviewed | Search scaffold | computer-science |
| 2022-02-16 |
[Deep reinforcement learning controls tokamak fusion plasma](/finding/2022-02-16-tokamak-plasma-control) |
Google DeepMind | DeepMind RL controller | Peer reviewed | Search scaffold | physics |
| 2021-12-01 |
[Two theorems found by machine pattern-spotting in knot theory and representation theory](/finding/2021-12-01-knot-theory-intuition) |
Google DeepMind (with Oxford and Sydney) | Supervised networks with gradient-based attribution | Peer reviewed | AI-assisted | mathematics |
| 2021-07-15 |
[Accurate protein structure prediction across the known proteome (AlphaFold2)](/finding/2021-07-15-alphafold2) |
Google DeepMind | AlphaFold2 | Independently checked | AI-led | biology |
| 2021-04-29 |
[Reinforcement learning refutes several conjectures in extremal combinatorics](/finding/2021-04-29-wagner-combinatorics) |
Tel Aviv University | Deep cross-entropy method (custom network) | Independently checked | Search scaffold | mathematics |
| 2020-02-20 |
[Halicin, an antibiotic found by a neural network screening a compound library](/finding/2020-02-20-halicin) |
MIT / Broad Institute | Directed message-passing graph neural network (Chemprop) | Peer reviewed | Search scaffold | medicine |
| 2017-12-14 |
[An eighth planet around Kepler-90 found by a neural network](/finding/2017-12-14-kepler-90i) |
Google Brain / University of Texas at Austin | Convolutional neural network (AstroNet) | Peer reviewed | Search scaffold | astronomy |