Free models forever | Master AI Automation in 4 hours Master AI Automation in 4 hours Course About Ayush Modules Sample chapter Toolbox The Microcap Minute Classroom / Module 04: APIs: How Software Talks / Chapter 6 Free models forever Watch first, then read. Same lesson, your pace. What you will learn – What OpenRouter is and why students should live there – Where free capacity hides (Gemini, :free models) – How to plug free models into terminal tools like OpenCode Two tools worth knowing by name OpenRouter (openrouter.ai), a single gateway to hundreds of models from every lab: OpenAI, Anthropic, Google, Alibaba (Qwen), Moonshot (Kimi), DeepSeek, Mistral and more. One account, one key, one bill (or no bill). Its magic for students: models tagged :free run at ₹0, real models like Qwen, DeepSeek and Llama variants, with daily limits instead of payment. OpenCode (opencode.ai), an open-source AI coding agent that lives in your terminal (the species you met in Module 3 Chapter 7). It reads files, writes code, runs commands, and connects to any provider, including every free OpenRouter model. Install when we reach Module 10; know today that it exists and eats free models happily. The free-first ladder Before paying anyone anything, climb this ladder: Free chat tiers , Kimi, Claude, ChatGPT, Qwen web apps: generous daily use for interactive work. Gemini free tier (Google AI Studio), famously large limits, huge context, multimodal: ideal student workhorse. OpenRouter :free models , programmatic access at ₹0 through your existing key; perfect for scripts, loops and experiments (paired with the off-peak habit from Chapter 5). Only then money , and when you do pay, OpenRouter’s metered pricing means you buy ₹50 of tokens, not a monthly subscription you forget to cancel. Rule of thumb: interactive thinking happens in chat apps; automated work runs on :free APIs; subscriptions enter only when something earns them monthly. Finding free models on OpenRouter On openrouter.ai → Models → filter price = FREE. Each listing shows context size and limits. Copy a model’s ID (e.g. deepseek/deepseek-chat-v3:free ) straight into Chapter 5’s script. Daily caps reset, combine two or three free models and your effective quota multiplies. Try it yourself Browse OpenRouter’s free list; pick three unfamiliar models. Point ask.py at each (one string change!) and ask all three: In two lines: what are you best at? Note the personality differences. Then check Gemini’s AI Studio free tier in a browser, upload a PDF and ask questions about it, free. Log findings in learn/free-stack.md . You’ve just audited your own zero-cost AI infrastructure. Key takeaways – OpenRouter = one key for hundreds of models; :free tags cost nothing. – OpenCode brings those free models into your terminal as a coding agent. – Free ladder: chat tiers → Gemini → :free APIs → paid last, metered not subscribed. – Rotating between free models multiplies your quota without spending a rupee. Download the exercise sheet (PDF) Module workbook (PDF) ← Prev: Talking to an AI with an API Next: Errors, rate limits and costs → Classroom / Module 04: APIs: How Software Talks / Chapter 6 Free models forever What you will learn – What OpenRouter is and why students should live there – Where free capacity hides (Gemini, :free models) – How to plug free models into terminal tools like OpenCode Two tools worth knowing by name OpenRouter (openrouter.ai), a single gateway to hundreds of models from every lab: OpenAI, Anthropic, Google, Alibaba (Qwen), Moonshot (Kimi), DeepSeek, Mistral and more. One account, one key, one bill (or no bill). Its magic for students: models tagged :free run at ₹0, real models like Qwen, DeepSeek and Llama variants, with daily limits instead of payment. OpenCode (opencode.ai), an open-source AI coding agent that lives in your terminal (the species you met in Module 3 Chapter 7). It reads files, writes code, runs commands, and connects to any provider, including every free OpenRou
Free models forever
Written by
in