Ginie — Smart Contract LLM

Generate, compile, and deploy smart contracts from plain English. Built for Solidity, Daml (Canton Network), and Rust (Vara Network).

Quickstart

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("GinieAI/Solidity-LLM")
model = AutoModelForCausalLM.from_pretrained(
    "GinieAI/Solidity-LLM",
    torch_dtype=torch.bfloat16,
    device_map="auto"
)

prompt = """### Instruction:
Write a Solidity ERC20 token with minting, burning, and owner controls.

### Response:
"""

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
    **inputs,
    max_new_tokens=800,
    temperature=0.7,
    do_sample=True,
    pad_token_id=tokenizer.eos_token_id
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

npm SDK

npm install ginie-sdk

import { Ginie } from 'ginie-sdk'

const ginie = new Ginie({ apiKey: 'your-key' })

const contract = await ginie.generate({
  prompt: 'ERC20 token with vesting schedule',
  chain: 'ethereum',
  audit: true
})

Model

Property	Value
Parameters	2B
Architecture	32 Transformer blocks
Context length	2048 tokens
Precision	bfloat16
Tokenizer	GPT-2
Base model	Chain-GPT/Solidity-LLM
License	MIT

Performance

Metric	Ginie v1	GPT-4o mini	DeepSeek-Coder-7B
Compilation rate	83%	78%	75%
OpenZeppelin compliance	65%	61%	58%
Gas efficiency	72%	65%	63%
Security score	58%	54%	51%

Evaluated on 100 prompts. Security via Slither static analysis.

Chains

Chain	Language	Status
Ethereum, Avalanche, Camp Network	Solidity	Live
Canton Network	Daml	v3
Vara Network	Rust	v3

Roadmap

Version	Ships
v1.0	Solidity — 2B params
v2.0	DISL + Zellic corpus — 7,800+ examples
v3.0	Daml + Rust support
v4.0	Weekly retraining flywheel on user data

Attribution

Fine-tuned from Chain-GPT/Solidity-LLM, itself based on Salesforce/codegen-2B-multi. MIT License.

ginie.xyz · Canton Network processes $6T in tokenised assets backed by Goldman Sachs, JPMorgan, and DTCC.

Generated contracts require expert review before production deployment.