This site lists neural networks whose weights and programs are stored fully on Ethereum mainnet. Each entry has a contract address and a live check that you can run from this page.

Models that store their files somewhere else, or run somewhere else and only post a proof, do not qualify. Each entry says where the model runs — inside Ethereum, or in your browser.

CENSUS

Registered onchain models
MODELAUTHORYEAR
陰XOR陽nahiko2021
ARAGNATIONDevi Parikh and Abhishek Das2023
INCHAINPEPEGANlifeonchain2023
CONCRETEHiggs × diid2025
I.ryley-o.eth2025
ARTIFICIAL AFTER ALLHan Saglam2026
AUTOMATE ATTENTIONdiid2026
SLONKSBlockhash2026
HELLO WORLD COMPUTERHan × 10012026
THE MARKET IS TALKINGFelixFelixFelix2026
remanencediid2026

11 entries. Each entry passed the automatic registry checks. Select an entry to run a new live check.

陰XOR陽

NAHIKO

陰XOR陽 facts
FIELDVALUE
TYPEtwo-layer neural network, 2-2-1 (MLP)
SIZE3 neurons
STORAGEfixed values in contract storage
OUTPUTthe XOR result as one boolean
ADDRESS0x2A648cb23D3ADf1BB171Fabc0119eA3eb951841f
YEAR2021

Three neurons in two layers compute XOR. The artist minted this work in August 2021 as a one-token NFT. The artist and the collector each set one input. The network turns the two inputs into yin, yang, or a blend. The weights are fixed values in the contract; anyone can read them. isYingYang() runs the model fully inside Ethereum.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

isYingYang() — runs the model on the two stored inputs and returns the XOR output as a bool

Requests go from your browser to public third-party Ethereum servers.

ARAGNATION

DEVI PARIKH AND ABHISHEK DAS

ARAGNATION facts
FIELDVALUE
TYPEcoordinate MLP (CPPN) — 5-32-32-32-32-32-32-3, ReLU
SIZE5,571 parameters (5,376 weights + 195 biases)
STORAGEdata contracts — plain number arrays in 7 locked script chunks on the Art Blocks core
OUTPUTan RGB color per (x, y) coordinate; some works use palette coloring instead
ADDRESS0x99a9B7c1116f9ceEB1652de04d5969CcE509B069
YEAR2023

Aragnation draws abstract compositions with a watercolor look. A small neural network learned to map each (x, y) position to a color. Its weights sit as plain numbers inside the project script, in seven locked data chunks on Ethereum. Your browser reads the script from Ethereum and runs it. Some works use palettes instead of the network.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

projectScriptByIndex(401, 3) — reads part 3 of 7 of the onchain program; this part is entirely model-weight numbers

Requests go from your browser to public third-party Ethereum servers.

INCHAINPEPEGAN

LIFEONCHAIN

INCHAINPEPEGAN facts
FIELDVALUE
TYPEGAN image generator (DCGAN-style)
SIZE≈2,900 parameters
STORAGEcontract storage; int16 fixed-point
TRAININGtrained offchain on pepe images, then written into the contract
OUTPUTthree 5×5 frames, built into an animated GIF inside Ethereum
ADDRESS0x65e9A5BbC1B26C0833B232355534cC404b0e69D4
YEAR2023

This GAN generator draws animated pepes inside Ethereum. A call to tokenURI() takes three seeds from the token id, runs the network three times, and builds an animated 5×5 GIF from the three frames. The artist trained the weights offchain and wrote them into contract storage. The weights, the network, and the GIF encoder all live at one address. The image exists nowhere else.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

tokenURI(1) — runs the generator three times (one per frame) and builds the animated GIF inside Ethereum

Requests go from your browser to public third-party Ethereum servers.

CONCRETE

HIGGS × DIID

CONCRETE facts
FIELDVALUE
TYPEGAN generator, first stage (single upsampling block)
SIZE79,680 weights
STORAGEdata contracts (SSTORE2); int16 fixed-point
OUTPUT12×8 RGB fingerprint, drawn as an SVG image inside Ethereum
ADDRESS0x4512759606f9E680361DD76C4c17F92031De7AcD
YEAR2025

The first collection of Higgs, an AI artist that diid built. The contract holds the first stage of the GAN generator; its weights live in data contracts on Ethereum. fingerprintSvg() runs the network inside Ethereum on a 16-byte seed and returns a 12×8 RGB fingerprint as an SVG image. The large display images are offchain; the check here covers only the onchain generator.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

fingerprintSvg(<token #1 seed>) — runs the model inside Ethereum; the call is too large for one public Ethereum server, so checks usually pass with 2 of 3 servers

Requests go from your browser to public third-party Ethereum servers.

I.

RYLEY-O.ETH

I. facts
FIELDVALUE
TYPEconditional VAE-GAN generator (cVAE-GAN)
SIZE297,859 parameters, quantized to uint8
STORAGEdata contracts — gzip shards in 24 locked contracts, 422,292 bytes total
OUTPUTone 64×64 RGB eye image per token, drawn in the browser
ADDRESS0x000000A35301Fa5784E820F489003FfcfFdc69a6
YEAR2025

I. draws 64×64 eyes with human expressions. A small cVAE-GAN generates each image. The full program — inference code and compressed weights — lives in 24 data contracts on Ethereum, locked forever. Your browser reads the program from Ethereum and runs it. No server and no outside library is needed.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

projectScriptByIndex(0, 5) — reads part 5 of 24 of the onchain program; parts 5–18 are the model weights (base64 text)

Requests go from your browser to public third-party Ethereum servers.

ARTIFICIAL AFTER ALL

HAN SAGLAM

ARTIFICIAL AFTER ALL facts
FIELDVALUE
TYPEcompressed decoder network
SIZE32-number latent, 4 conv layers; one network in each of 256 contracts
STORAGEeach contract's own bytecode; int8, Huffman-compressed
OUTPUT32×32 grayscale image, 1,024 bytes
ADDRESS0xbDF5A7F5e5AE18eE580139297C63a89fc1d18B5f
YEAR2026

This work is 256 contracts on Ethereum. Each contract holds one seed and one compressed neural network in its own code. A call to artificial() runs the network inside Ethereum and returns a 32×32 grayscale image (1024 bytes). The contract computes the image at call time and stores no image. The address here is one example piece; a factory deployed all 256.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

artificial() — returns a 32×32 image as 1024 pixel bytes (1088 raw bytes with ABI framing)

Requests go from your browser to public third-party Ethereum servers.

AUTOMATE ATTENTION

DIID

AUTOMATE ATTENTION facts
FIELDVALUE
TYPE5-layer transformer language model (Llama-style)
SIZEdim 48, hidden 128, 8 heads, 2,048-token vocabulary
STORAGEdata contracts; int8 quantized
OUTPUTone aphorism per NFT, drawn as an SVG image inside Ethereum
ADDRESS0x5F8E7D750E75b44747C058A204D8DEa0D18fA5d3
YEAR2026

This contract is a complete transformer language model that runs inside Ethereum. Data contracts on Ethereum store its weights. A companion NFT contract calls the model to write one permanent aphorism for each token. The model also draws the text as an SVG image inside Ethereum. The address here is the language model itself.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

logits([1]) — runs the model once and returns int16 scores (logits) for all 2048 vocabulary tokens

Requests go from your browser to public third-party Ethereum servers.

SLONKS

BLOCKHASH

SLONKS facts
FIELDVALUE
TYPEcandidate-palette classifier — per-pixel 18-way argmax over learned embeddings
SIZE203,680 learned int8 values (10,000×10 embeddings + 576×18×10 pixel heads)
STORAGEdata contracts (SSTORE2) — 214,048 int8 bytes in 9 locked contracts
TRAININGtrained offchain on the CryptoPunks image set; 95.77% pixel accuracy
OUTPUT576 palette indexes — a 24×24 reconstruction of the source punk
ADDRESS0xCa116243a2013ED33015c776ee37310b199Ee80c
YEAR2026

Slonks redraws all 10,000 CryptoPunks with a small model that lives on Ethereum. For each pixel, the model scores 18 candidate colors against a learned 10-byte embedding. The best score wins. The weights fill nine data contracts and can never change. The model runs inside Ethereum. The pixels the model gets wrong are the art.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

renderSourcePixels(0) — runs the model inside Ethereum for punk #0 and returns 576 palette indexes, one per pixel

Requests go from your browser to public third-party Ethereum servers.

HELLO WORLD COMPUTER

HAN × 1001

HELLO WORLD COMPUTER facts
FIELDVALUE
TYPEsingle-layer perceptron
SIZE1,024 weights (hashed text n-grams)
STORAGEcontract storage; int32 packed in uint256[128]
TRAININGlearns inside Ethereum — weights start at zero and update when each question settles
OUTPUTa yes/no label with a confidence value
ADDRESS0x6901afceb66564a9b6E7193561fB78F0878b5906
YEAR2026

This perceptron lives inside one contract. The contract stores the weights, trains them, and answers questions. You send it a short text (up to 64 bytes). ask() returns a yes/no label and a confidence value. All weights started at zero. Anyone can post a yes/no question for a fee. Each settled question updates the weights inside the contract.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

ask("hello world") — runs the model on the text and returns (label uint8, confidence int256)

Requests go from your browser to public third-party Ethereum servers.

THE MARKET IS TALKING

FELIXFELIXFELIX

THE MARKET IS TALKING facts
FIELDVALUE
TYPEtwo-layer echo-state network
SIZE20 hidden neurons, 4 inputs
STORAGEderived at read time — each block's weights come from live market activity, not stored
OUTPUTone deterministic sentence per block
ADDRESS0xd6D8555E131c0C431601aF62141A61a9b1aa6b13
YEAR2026

This network reads the market itself. Its hidden layers are two baskets of live token markets. Each block, it derives fresh weights from the trade activity of those tokens. Ethereum stores the renderer program; your browser runs it. The output is one sentence per block, from a 299-word vocabulary. stateAt() returns the exact state at any past block, so anyone can compute each sentence again.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

stateAt(25599744) — returns the saved inputs and mood state; the renderer rebuilds the network from these

Requests go from your browser to public third-party Ethereum servers.

remanence

DIID

remanence facts
FIELDVALUE
TYPEconvolutional image decoder (4 conv layers, two output heads)
SIZE1,512 conv weights (int16) plus 110,592 stored latent values
STORAGEdata contracts (SSTORE2) — weights, activation tables, and 16 painting latents
OUTPUTone 192×144 indexed-color BMP, computed inside the EVM
ADDRESS0x788Ee0B4Fb97657Cc2F35829d19C0f129c349B5a
YEAR2026

remanence stores sixteen paintings on Ethereum as compressed fragments. A small neural decoder turns fragments back into a picture. The decoder weights and the fragments are in data contracts. The model runs inside Ethereum: one read-only call runs every layer and returns a finished image. Collectors combine fragments to restore a painting, or mix paintings into a chimera.

VERIFICATION

The check reads the contract code and runs the call on public Ethereum servers.

tokenURI(101) — runs the decoder in EVM opcodes on token 101's fragments and returns the image as token metadata

Requests go from your browser to public third-party Ethereum servers.

METHOD

Software checks the schema, file name, evidence links, and declared return size of each entry. Two or more public Ethereum servers must then agree on the contract code and the exact call result at one block. These checks prove that the call is reproducible. They do not prove that the code is the described neural network. A person also reviews the source evidence. To correct or add an entry, open a pull request.