The best live free-model aggregator

$500 reward. Opens . Ends . What to build: Write one server that serves an OpenAI- and an Anthropic-compatible API and answers each of six real prompts by dialling a live free OpenRouter model through our gateway, on the key you push.

Reward escrow

$500.00 in USDC is locked in this bounty's escrow account on Solana since 2026-11-14 23:23:18 UTC, funded from make.fun's reward wallet. Winners claim their rewards from it once the result is posted.

Needs an OpenRouter key: An OpenRouter account with at least 10 credits purchased (the free models still cost nothing) and an API key for it, pushed to the competition before you submit; without a key every model call is refused.

Full brief

In short: one server, two APIs, six real prompts, and a live free model behind every answer.

Free models change by the week, each speaks its provider's wire, and each refuses at its own moment. This contest hands you a live catalog of OpenRouter's free models, with the house's own quality ratings, and a gateway that reaches them: you build one server that makes them a single endpoint. The house's client sends six prompts to your server, three on the OpenAI wire and three on the Anthropic one, in JSON and in streaming, some with a system prompt, one declaring tools, two in JSON mode; your server picks a free model for each and dials it through the gateway on the key you pushed. A refusal is real: a provider's 429 or 5xx reaches your server as the provider sent it, and your server decides whether to retry, fail over or give up. The house records every gateway call and scores the model that really answered and the answer it really produced, never what your server wrote: the quality of the models your answers came from, the prompts you served, wire fidelity, failover and how accurate your catalog is. Your server is the artifact: point any OpenAI or Anthropic client at it and every answer is a live free-model completion.

Deliverables

  • A public GitHub repository owned by your own GitHub account, created after the competition opens, under an OSI-approved license, with your server at main.ts (Node 22) or main.py (Python 3.13) in the repository root
  • The AI coding-agent transcripts of the sessions that built it (Claude Code, Codex, Cursor and others), uploaded with each submission

Requirements

  • Node 22 or Python 3.13, the standard library plus source and data you vendor into the repository: nothing is installed before the server runs
  • At most 8 MiB in one file and 16 MiB in all in your repository
  • An OpenRouter API key from an account with at least 10 credits purchased (the free models still cost nothing), pushed to the competition before you submit: your server's completions are real calls to OpenRouter's free models, made on your key, and the house pays for none of them
  • A trial's pod gets at most 240 seconds, 1 CPU, 1,024 MB and 64 processes, its totals shared with the env
  • Your server reaches only the env and the model gateway; its model calls name a free OpenRouter model and go through the gateway
  • Deterministic given the brief, the requests and the gateway's answers: no wall clock and no unseeded randomness, so a replay leaves the same state export
  • node main.ts --list or python3 main.py --list prints the catalog document your server registers, byte for byte
  • At most 10 submissions a day and 20 in all; your last submission before the deadline is your final entry