OpenGradient · adambalogh · Apr 15, 2026 · Apr 10, 2026 · Apr 10, 2026 · Apr 15, 2026
diff --git a/CLAUDE.md b/CLAUDE.md
@@ -99,7 +99,7 @@ make chat-tool
 - **Dual Authentication**: Model Hub (email/password) + blockchain (private key)
 - **x402 Payments**: Streamed micropayment protocol for LLM inference
 - **Digital Twins**: Chat integration with twin.fun personas
-- **Dual Chain**: Base Sepolia (LLM) + OpenGradient testnet (Alpha on-chain inference)
+- **Dual Chain**: Base mainnet (LLM) + OpenGradient testnet (Alpha on-chain inference)
 
 ## Configuration
 

diff --git a/README.md b/README.md
@@ -52,10 +52,9 @@ For current network RPC endpoints, contract addresses, and deployment informatio
 
 Before using the SDK, you will need:
 
-1. **Private Key**: An Ethereum-compatible wallet private key funded with **Base Sepolia OPG tokens** for x402 LLM payments
-2. **Test Tokens**: Obtain free test tokens from the [OpenGradient Faucet](https://faucet.opengradient.ai) for testnet LLM inference
-3. **Alpha Testnet Key** (Optional): A private key funded with **OpenGradient testnet gas tokens** for Alpha Testnet on-chain inference (can be the same or a different key)
-4. **Model Hub Account** (Optional): Required only for model uploads. Register at [hub.opengradient.ai/signup](https://hub.opengradient.ai/signup)
+1. **Private Key**: An Ethereum-compatible wallet private key funded with **Base OPG tokens** for x402 LLM payments
+2. **Alpha Testnet Key** (Optional): A private key funded with **OpenGradient testnet gas tokens** for Alpha Testnet on-chain inference (can be the same or a different key)
+3. **Model Hub Account** (Optional): Required only for model uploads. Register at [hub.opengradient.ai/signup](https://hub.opengradient.ai/signup)
 
 ### Configuration
 
@@ -87,7 +86,7 @@ The SDK provides separate clients for each service. Create only the ones you nee
 import os
 import opengradient as og
 
-# LLM inference — settles via x402 on Base Sepolia using OPG tokens
+# LLM inference — settles via x402 on Base using OPG tokens
 llm = og.LLM(private_key=os.environ.get("OG_PRIVATE_KEY"))
 
 # Alpha Testnet — on-chain inference on the OpenGradient network using testnet gas tokens

diff --git a/docs/CLAUDE_SDK_USERS.md b/docs/CLAUDE_SDK_USERS.md
@@ -40,7 +40,7 @@ print(result.chat_output["content"])
 Each service has its own client class:
 
 ```python
-# LLM inference (Base Sepolia OPG tokens for x402 payments)
+# LLM inference (Base OPG tokens for x402 payments)
 llm = og.LLM(private_key="0x...")
 
 # Connect directly to a known TEE IP instead of using the on-chain registry.

diff --git a/docs/opengradient/client/index.md b/docs/opengradient/client/index.md
@@ -10,7 +10,7 @@ OpenGradient Client -- service modules for the SDK.
 
 ## Modules
 
-- **[llm](./llm)** -- LLM chat and text completion with TEE-verified execution and x402 payment settlement (Base Sepolia OPG tokens)
+- **[llm](./llm)** -- LLM chat and text completion with TEE-verified execution and x402 payment settlement (Base OPG tokens)
 - **[model_hub](./model_hub)** -- Model repository management: create, version, and upload ML models
 - **[alpha](./alpha)** -- Alpha Testnet features: on-chain ONNX model inference (VANILLA, TEE, ZKML modes), workflow deployment, and scheduled ML model execution (OpenGradient testnet gas tokens)
 - **[twins](./twins)** -- Digital twins chat via OpenGradient verifiable inference
@@ -22,7 +22,7 @@ OpenGradient Client -- service modules for the SDK.
 ```python
 import opengradient as og
 
-# LLM inference (Base Sepolia OPG tokens)
+# LLM inference (Base OPG tokens)
 llm = og.LLM(private_key="0x...")
 llm.ensure_opg_approval(min_allowance=5)
 result = await llm.chat(model=og.TEE_LLM.CLAUDE_HAIKU_4_5, messages=[...])

diff --git a/docs/opengradient/client/llm.md b/docs/opengradient/client/llm.md
@@ -231,11 +231,11 @@ Permit2ApprovalResult: Contains ``allowance_before``,
 
 ### Automatic Endpoint Discovery (Production)
 
-By default, `LLM()` constructor automatically discovers active TEE endpoints from the on-chain TEE registry using the `rpc_url` parameter (defaults to the Base Sepolia testnet).
+By default, `LLM()` constructor automatically discovers active TEE endpoints from the on-chain TEE registry using the `rpc_url` parameter (defaults to the OpenGradient devnet).
 
 **Key Points:**
 - The TEE endpoint is **dynamically discovered** from the registry
-- x402 payments are always settled on **Base Sepolia**, regardless of which TEE endpoint serves your request
+- x402 payments are always settled on **Base**, regardless of which TEE endpoint serves your request
 - This is the recommended approach for production use
 
 **Example:**
@@ -250,7 +250,7 @@ For development, testing, or self-hosted TEE servers, use `LLM.from_url()` with
 
 **Key Points:**
 - TLS certificate verification is disabled (suitable for self-signed certs)
-- x402 payment settlement still occurs on Base Sepolia
+- x402 payment settlement still occurs on Base
 - Intended for non-production environments only
 
 **Example:**
@@ -264,5 +264,5 @@ llm = og.LLM.from_url(
 
 ### Important: x402 Payment Settlement
 
-Regardless of which TEE endpoint serves your inference request, **x402 payment settlement always occurs on Base Sepolia blockchain**. This ensures all payments are recorded on-chain for auditability and ensures consistent settlement across all TEE providers.
+Regardless of which TEE endpoint serves your inference request, **x402 payment settlement always occurs on Base blockchain**. This ensures all payments are recorded on-chain for auditability and ensures consistent settlement across all TEE providers.
 
diff --git a/docs/opengradient/index.md b/docs/opengradient/index.md
@@ -19,7 +19,7 @@ inference was performed correctly.
 
 The SDK operates across two chains with separate private keys:
 
-- **[llm](./client/llm)** (``og.LLM``) -- LLM chat and completion with TEE-verified execution. Pays via x402 on **Base Sepolia** (requires OPG tokens).
+- **[llm](./client/llm)** (``og.LLM``) -- LLM chat and completion with TEE-verified execution. Pays via x402 on **Base** (requires OPG tokens).
 - **[alpha](./client/alpha)** (``og.Alpha``) -- On-chain ONNX model inference with VANILLA, TEE, or ZKML verification. Pays gas on the **OpenGradient alpha testnet**.
 - **[model_hub](./client/model_hub)** (``og.ModelHub``) -- Model repository management: create, version, and upload ML models. Requires email/password auth.
 - **[twins](./client/twins)** (``og.Twins``) -- Digital twins chat via verifiable inference. Requires a twins API key.

diff --git a/examples/README.md b/examples/README.md
@@ -8,7 +8,7 @@ Before running any examples, ensure you have:
 
 1. **Installed the SDK**: `pip install opengradient`
 2. **Set up your credentials**: Configure your OpenGradient account using environment variables:
-   - `OG_PRIVATE_KEY`: Private key funded with **Base Sepolia OPG tokens** for x402 LLM payments (can be obtained from our [faucet](https://faucet.opengradient.ai)). Also used for Alpha Testnet on-chain inference (requires **OpenGradient testnet gas tokens**).
+   - `OG_PRIVATE_KEY`: Private key funded with **Base OPG tokens** for x402 LLM payments.
    - `OG_MODEL_HUB_EMAIL`: (Optional) Your Model Hub email for model uploads
    - `OG_MODEL_HUB_PASSWORD`: (Optional) Your Model Hub password for model uploads
 
@@ -140,7 +140,7 @@ Each sub-client is created independently with the credentials it needs:
 import os
 import opengradient as og
 
-# LLM inference (Base Sepolia OPG tokens for x402 payments)
+# LLM inference (Base OPG tokens for x402 payments)
 llm = og.LLM(private_key=os.environ.get("OG_PRIVATE_KEY"))
 
 # On-chain model inference (OpenGradient testnet gas tokens)

diff --git a/integrationtest/README.md b/integrationtest/README.md
@@ -6,12 +6,12 @@ End-to-end tests that exercise the OpenGradient SDK against live services.
 
 ### LLM (`llm/`)
 
-Tests LLM chat and streaming chat via the x402 payment flow on Base Sepolia.
+Tests LLM chat and streaming chat via the x402 payment flow on Base.
 
 Each run creates a **fresh Ethereum account**, funds it with ETH (for gas) and OPG tokens from a funder wallet, approves Permit2, and then runs chat requests against a TEE-verified model.
 
 **Requirements:**
-- `PRIVATE_KEY` env var — private key of a funded wallet on Base Sepolia that holds OPG tokens.
+- `PRIVATE_KEY` env var — private key of a funded wallet on Base that holds OPG tokens.
 
 ```bash
 make llm_integrationtest

diff --git a/integrationtest/llm/test_llm_chat.py b/integrationtest/llm/test_llm_chat.py
@@ -6,7 +6,7 @@
 from web3 import Web3
 
 import opengradient as og
-from opengradient.client.opg_token import BASE_OPG_ADDRESS, BASE_SEPOLIA_RPC
+from opengradient.client.opg_token import BASE_OPG_ADDRESS, BASE_MAINNET_RPC
 
 # Minimal ERC20 ABI for transfer
 ERC20_TRANSFER_ABI = [
@@ -37,7 +37,7 @@
 
 def _fund_account(funder_key: str, recipient_address: str):
     """Transfer ETH (for gas) and OPG tokens from the funder to the recipient."""
-    w3 = Web3(Web3.HTTPProvider(BASE_SEPOLIA_RPC))
+    w3 = Web3(Web3.HTTPProvider(BASE_MAINNET_RPC))
     funder = Account.from_key(funder_key)
     funder_addr = Web3.to_checksum_address(funder.address)
     recipient = Web3.to_checksum_address(recipient_address)

diff --git a/src/opengradient/__init__.py b/src/opengradient/__init__.py
@@ -10,7 +10,7 @@
 
 The SDK operates across two chains with separate private keys:
 
-- **`opengradient.client.llm`** (``og.LLM``) -- LLM chat and completion with TEE-verified execution. Pays via x402 on **Base Sepolia** (requires OPG tokens).
+- **`opengradient.client.llm`** (``og.LLM``) -- LLM chat and completion with TEE-verified execution. Pays via x402 on **Base** (requires OPG tokens).
 - **`opengradient.client.alpha`** (``og.Alpha``) -- On-chain ONNX model inference with VANILLA, TEE, or ZKML verification. Pays gas on the **OpenGradient alpha testnet**.
 - **`opengradient.client.model_hub`** (``og.ModelHub``) -- Model repository management: create, version, and upload ML models. Requires email/password auth.
 - **`opengradient.client.twins`** (``og.Twins``) -- Digital twins chat via verifiable inference. Requires a twins API key.

diff --git a/src/opengradient/client/__init__.py b/src/opengradient/client/__init__.py
@@ -3,7 +3,7 @@
 
 ## Modules
 
-- **`opengradient.client.llm`** -- LLM chat and text completion with TEE-verified execution and x402 payment settlement (Base Sepolia OPG tokens)
+- **`opengradient.client.llm`** -- LLM chat and text completion with TEE-verified execution and x402 payment settlement (Base OPG tokens)
 - **`opengradient.client.model_hub`** -- Model repository management: create, version, and upload ML models
 - **`opengradient.client.alpha`** -- Alpha Testnet features: on-chain ONNX model inference (VANILLA, TEE, ZKML modes), workflow deployment, and scheduled ML model execution (OpenGradient testnet gas tokens)
 - **`opengradient.client.twins`** -- Digital twins chat via OpenGradient verifiable inference
@@ -15,7 +15,7 @@
 ```python
 import opengradient as og
 
-# LLM inference (Base Sepolia OPG tokens)
+# LLM inference (Base OPG tokens)
 llm = og.LLM(private_key="0x...")
 llm.ensure_opg_approval(min_allowance=5)
 result = await llm.chat(model=og.TEE_LLM.CLAUDE_HAIKU_4_5, messages=[...])

diff --git a/src/opengradient/client/llm.py b/src/opengradient/client/llm.py
@@ -28,14 +28,13 @@
 
 X402_PROCESSING_HASH_HEADER = "x-processing-hash"
 X402_PLACEHOLDER_API_KEY = "0x1234567890abcdef1234567890abcdef1234567890abcdef1234567890abcdef"
-BASE_TESTNET_NETWORK = "eip155:84532"
-BASE_TESTNET_RPC = os.getenv("BASE_TESTNET_RPC", "https://sepolia.base.org")
+BASE_MAINNET_NETWORK = "eip155:8453"
+BASE_MAINNET_RPC = os.getenv("BASE_MAINNET_RPC", "https://base-rpc.publicnode.com")
 
 _CHAT_ENDPOINT = "/v1/chat/completions"
 _COMPLETION_ENDPOINT = "/v1/completions"
 _REQUEST_TIMEOUT = 60
 
-
 @dataclass(frozen=True)
 class _ChatParams:
     """Bundles the common parameters for chat/completion requests."""
@@ -87,7 +86,7 @@ def __init__(
             raise ValueError("A private key is required to use the LLM client. Pass a valid private_key to the constructor.")
         self._wallet_account: LocalAccount = Account.from_key(private_key)
 
-        x402_client = LLM._build_x402_client(private_key, rpc_url=BASE_TESTNET_RPC)
+        x402_client = LLM._build_x402_client(private_key, rpc_url=BASE_MAINNET_RPC)
         onchain_registry = TEERegistry(rpc_url=rpc_url, registry_address=tee_registry_address)
         self._tee: TEEConnectionInterface = RegistryTEEConnection(x402_client=x402_client, registry=onchain_registry)
 
@@ -117,16 +116,16 @@ def from_url(
         return instance
 
     @staticmethod
-    def _build_x402_client(private_key: str, rpc_url: str = BASE_TESTNET_RPC) -> x402Client:
+    def _build_x402_client(private_key: str, rpc_url: str = BASE_MAINNET_RPC) -> x402Client:
         """Build the x402 payment stack from a private key."""
         account = Account.from_key(private_key)
         signer = EthAccountSigner(account)
         client = x402Client()
-        register_exact_evm_client(client, signer, networks=[BASE_TESTNET_NETWORK])
+        register_exact_evm_client(client, signer, networks=[BASE_MAINNET_NETWORK])
         register_upto_evm_client(
             client,
             signer,
-            networks=[BASE_TESTNET_NETWORK],
+            networks=[BASE_MAINNET_NETWORK],
             rpc_url=rpc_url,
         )
         return client

diff --git a/src/opengradient/client/opg_token.py b/src/opengradient/client/opg_token.py
@@ -1,5 +1,6 @@
 """OPG token Permit2 approval utilities for x402 payments."""
 
+import os
 import logging
 import time
 from dataclasses import dataclass
@@ -12,8 +13,8 @@
 
 logger = logging.getLogger(__name__)
 
-BASE_OPG_ADDRESS = "0x240b09731D96979f50B2C649C9CE10FcF9C7987F"
-BASE_SEPOLIA_RPC = "https://sepolia.base.org"
+BASE_OPG_ADDRESS = "0xFbC2051AE2265686a469421b2C5A2D5462FbF5eB"
+BASE_MAINNET_RPC = os.getenv("BASE_MAINNET_RPC", "https://base-rpc.publicnode.com")
 APPROVAL_TX_TIMEOUT = 120
 ALLOWANCE_CONFIRMATION_TIMEOUT = 120
 ALLOWANCE_POLL_INTERVAL = 1.0
@@ -138,7 +139,7 @@ def _send_approve_tx(
 
 def _get_web3_and_contract():
     """Create a Web3 instance and OPG token contract."""
-    w3 = Web3(Web3.HTTPProvider(BASE_SEPOLIA_RPC))
+    w3 = Web3(Web3.HTTPProvider(BASE_MAINNET_RPC))
     token = w3.eth.contract(address=Web3.to_checksum_address(BASE_OPG_ADDRESS), abi=ERC20_ABI)
     spender = Web3.to_checksum_address(PERMIT2_ADDRESS)
     return w3, token, spender

diff --git a/tests/opg_token_test.py b/tests/opg_token_test.py
@@ -54,7 +54,7 @@ def _setup_approval_mocks(mock_web3, mock_wallet, contract):
 
     mock_web3.eth.get_transaction_count.return_value = 7
     mock_web3.eth.gas_price = 1_000_000_000
-    mock_web3.eth.chain_id = 84532
+    mock_web3.eth.chain_id = 8453
 
     signed = MagicMock()
     signed.raw_transaction = b"\x00"

diff --git a/tutorials/01-verifiable-ai-agent.md b/tutorials/01-verifiable-ai-agent.md
@@ -29,9 +29,8 @@ Ethereum private key works -- you can generate one with any Ethereum wallet.
 export OG_PRIVATE_KEY="0x..."
 ```
 
-> **Faucet:** Get free OPG tokens on Base Sepolia at <https://faucet.opengradient.ai/>
-> so your wallet can pay for inference transactions. All x402 LLM payments currently
-> settle on Base Sepolia.
+>All x402 LLM payments currently
+> settle on Base.
 
 ## Step 1: Initialize and Create the LangChain Adapter
 

diff --git a/tutorials/02-streaming-multi-provider.md b/tutorials/02-streaming-multi-provider.md
@@ -27,10 +27,8 @@ Export your OpenGradient private key:
 export OG_PRIVATE_KEY="0x..."
 ```
 
-> **Faucet:** Get free OPG tokens on Base Sepolia at https://faucet.opengradient.ai/
->
-> All x402 LLM payments currently settle on Base Sepolia using OPG tokens. If you see
-> x402 payment errors, make sure your wallet has sufficient OPG on Base Sepolia.
+
+> All x402 LLM payments currently settle on Base using OPG tokens.
 
 ## Step 1: Basic Non-Streaming Chat
 

diff --git a/tutorials/03-verified-tool-calling.md b/tutorials/03-verified-tool-calling.md
@@ -28,10 +28,7 @@ You need an OpenGradient private key funded with test tokens:
 export OG_PRIVATE_KEY="0x..."
 ```
 
-> **Faucet:** Get free OPG tokens on Base Sepolia at https://faucet.opengradient.ai/
->
-> All x402 LLM payments currently settle on Base Sepolia using OPG tokens. If you see
-> x402 payment errors, make sure your wallet has sufficient OPG on Base Sepolia.
+> All x402 LLM payments currently settle on Base using OPG tokens.
 
 ## Step 1: Initialize the Client