Opper AI

Opper AI is an AI evaluation platform and gateway that benchmarks decision models using a Pac-Man game.

Opper AI provides a Pac-Man benchmark called jevman to test AI decision models in real time against classic arcade ghosts. Six models played 100 games each, and users can view the leaderboard, watch recorded games, or play against the models. The benchmark is open source and accepts any model behind an HTTP endpoint that can receive maze JSON and return movement directions. Additionally, Opper offers an AI gateway providing access to over 700 models from more than 50 providers at provider rates without token markups. The platform features pay-as-you-go credit purchases with a 3 percent fee, optional control plane features for a 5.5 percent fee, and OpenAI SDK compatibility for easy migration.

Opper AI at a glance

Best for: Software engineer
Runs on: Web, API, self-hosted, open source
Pricing: Pay-as-you-go credits; 3 percent fee on top-ups for Gateway; 5.5 percent fee with Control Plane; Enterprise plans available.
Released: July 2025
Category: AI agent tool
Website: opper.ai

Key features

Pac-Man benchmark for decision models
Open source benchmark repository
HTTP endpoint model integration
Leaderboard with mean scores and margins of error
Gateway with 700 plus models from 50 plus providers
Pay-as-you-go credit system with no token markup
OpenAI SDK compatibility for migration
Optional control plane with spend limits and tracing

Use it for

Benchmark decision models using a Pac-Man game
Access multiple AI models through a single gateway
Migrate existing applications from OpenAI or Anthropic
Route requests with fallback model lists

Alternatives to Opper AI

OpenSquilla — Maximizes AI intelligence density while minimizing token costs through smart routing.
Composio — Equips AI agents with tools, managed user auth, and sandboxes across 1,500+ apps
Catan Advisor — Measures winning strategies across thousands of simulated games.
OpenViking — Unifies agent memory, knowledge RAG, and skills into a single directory structure.
OpenBot — Runs entirely on your own infrastructure with strict document permissions and custom.
Kungfu — Maintains work state and context when switching between different AI agents.
Rakazo — Complete ownership of data, models, and execution environment.
OpenHarness — Manages Claude Code, Codex, and Hermes from a single command center.
FastGPT — It uses hybrid retrieval and auto-cleansed data to prevent hallucinations.
OpenAgent — Self-hosted single binary with MCP support, 30+ model providers, and zero data leakage.
APIStack — Connect AI agents to over 100 tools using one API key or a single MCP endpoint.
OpenBot — Runs AI agents locally using existing subscriptions or local models.

All Opper AI alternatives