Knowledge Base Sections ▾

Navigation

▸ Start here By roles

Categories

Tools 54
Glossary 12

Tools

Cursor + Gonka AI - cheap LLM for coding

Cursor is one of the most popular AI code editors of 2026. Autocomplete, refactoring, test generation, code explanation — all of it runs on LLM models. But OpenAI and Anthropic API costs can reach $30 per 1M tokens, and with active development your monthly bill easily exceeds $100.

JoinGonka Gateway is an OpenAI-compatible API to the decentralized Gonka network at $0.0062/1M tokens. That is about 810 times cheaper than OpenAI GPT-5.5. The network's models — MiniMax M2.7, DeepSeek V4 Flash and GLM-5.3 Flash, all built on the MoE architecture — deliver quality comparable to GPT-5.5 and Claude Sonnet 4.6, but for pennies.

Connecting to Cursor takes 30 seconds. Here's a step-by-step guide.

Step 1: Get JoinGonka API Key

To connect Cursor to Gonka, you need an API key from the JoinGonka Gateway — our gateway to the decentralized network.

  1. Go to gate.joingonka.ai/register and create an account.
  2. After confirming your address, you'll receive a bonus of 3M free tokens — enough for thousands of AI requests.
  3. In your Dashboard, go to the API Keys section and click Create Key.
  4. Copy the key — it starts with jg-, for example: jg-abc123def456.

The key works with any OpenAI-compatible client: Claude Code, Aider, Continue.dev, and of course Cursor.

Security: the key is shown only once at creation. Store it in a safe place. If you lose it, create a new one and deactivate the old one in the Dashboard.

Step 2: Configure Cursor

Open Cursor and go to settings:

  1. Settings → Models (or Ctrl+, → Models)
  2. In the OpenAI API Key field, paste your jg-xxx key.
  3. Enable Override OpenAI Base URL and enter: https://gate.joingonka.ai/v1
  4. In the model list, add: MiniMaxAI/MiniMax-M2.7
  5. Select this model in the chat and agent model switcher.

Check: open Cursor Chat (Ctrl+L) and type any request, for example: "Write a sorting function in Python." If you get a response, everything is set up correctly.

Optimization tips:

  • Use the same model for Chat and Composer — MiniMax M2.7 understands code context well; switch to GLM-5.3 Flash for tangled logic, and to DeepSeek V4 Flash for large repositories.
  • Enable streaming for faster response — the answer starts appearing immediately without waiting for full generation.
  • If you need a different response format, use the System Prompt in Cursor settings for fine-tuning.

Possible errors:

  • 401 Unauthorized — check that the key starts with jg- and is active in the Dashboard.
  • 404 Not Found — make sure the Base URL is exactly https://gate.joingonka.ai/v1 (with /v1 at the end).
  • Timeout — the first request may take 5-10 seconds due to a cold start of a node in the network. Subsequent ones are faster.
  • The built-in Cursor model stopped responding — while the Override OpenAI Base URL toggle is on, Cursor sends all OpenAI-family requests to your endpoint, including those to its own built-in models. Turn the toggle off while working with Cursor models and turn it back on for network models.

Note: Chat and Agent mode are routed to your endpoint; Cursor Tab autocomplete runs on Cursor's own models and does not go through a third-party API — this is a limitation of Cursor itself, not the gateway. On Cursor team plans (Teams and Enterprise), requests with your own key incur Cursor's own rate — $0.25 per 1M tokens (Cursor documentation, October 2026); on individual plans, requests with your own key are not counted toward Cursor usage.

Tip. The command npx @joingonka/setup --tool cursor will print ready-to-use values for these fields — base URL, key, model id and its real limits: context window and response ceiling — and will verify with a live request that the gateway accepts them. Cursor itself is configured only through the UI — it has no file that can be safely edited, so values must be pasted manually.

Cost Comparison

Cursor is a heavy token user: Chat, Composer, agent edits, refactoring — all of these are LLM requests. In typical development (4-6 hours a day), 3-10M tokens are consumed daily. Let's compare costs:

ProviderModelPrice per 1MDay (~7M)Month (~150M)
JoinGonka GatewayMiniMax M2.7 / DeepSeek V4 Flash / GLM-5.3 Flash$0.0099$0.069$1.49
OpenAIGPT-5.5~$5.00$35$750
AnthropicClaude Sonnet 4.6~$7.00$49$1,050

Savings: over a month of active development, JoinGonka Gateway will save $750—1,000 compared to OpenAI/Anthropic. For a team of 5 developers — $3,750—5,250 per month.

At the same time, the quality of the network's model responses is comparable to GPT-5.5 and Claude on coding tasks: function generation, refactoring, writing tests, explaining code. They all support native tool calling, which makes them suitable for agent scenarios as well.

Model parameters: MiniMax M2.7 supports a 180K token context window; DeepSeek V4 Flash and GLM-5.3 Flash have longer ones. Maximum response length through the Gateway is 8192 tokens for MiniMax M2.7 and GLM-5.3 Flash, 32768 for DeepSeek V4 Flash — more than enough for chat and code edits. List of available models: GET https://gate.joingonka.ai/v1/models.

Cursor + Gonka = a full-fledged AI coding assistant for pennies. 30 seconds to set up, free starting tokens, $0.50 per month instead of $750+ with OpenAI. The network's model quality is comparable to GPT-5.5 on coding tasks.

Want to learn more?

Explore other sections or start earning GNK right now.

Get free tokens →