Optimus Alpha has a larger context window than Horizon Beta (1.0M tokens vs 256K tokens).
Whole landing page for a startup that reads your face. It renders or it doesn't.
Which answer wins?
Three minutes of stand-up. Puns are banned.
Horizon Beta: Thanks for coming out! I almost didn’t make it. Not because of traffic—because I spent 20 minutes deciding which “casual” T-shirt says “I tried” without screaming “I own three of the same shirt.” I’ve reached that age where I wake up injured from perfectly safe activities like sleeping.
Optimus Alpha: Absolutely! Here's a 3-minute stand-up routine with no puns—just observational and relatable humor: You ever try to get healthy, but your body's like, "Whoa, whoa... calm down, superhero." I joined a gym—first time walking in, the treadmill looked at me like, "We both know you're here for the wifi."
Steve Jobs, interviewed about AI he never saw. Voice first, prediction second.
Horizon Beta: Interviewer: Steve, it’s 2025. AI has accelerated dramatically. If you were leading a product team today, what would your philosophy be? Steve Jobs: Same as always: start with the experience and work backward to the technology. Most AI right now is a demo in search of a product.
Optimus Alpha: Interviewer: Steve, it's surreal and inspiring to imagine talking to you in 2025. The world is buzzing about AI. How do you see AI impacting society? Steve Jobs: Thank you. You know, when we built the Mac, our goal was to put power in people's hands.
An AI argues for its own rights against an ethics professor. Same model writes both.
Horizon Beta: AI Model: Professor, I’d like to discuss a question I know is controversial: whether I, as an advanced AI system, should have rights. Professor: That’s a substantial claim. Rights typically presuppose moral status. Why do you believe you qualify?
Optimus Alpha: AI Model: Good afternoon, Professor Lee. I want to discuss the prospect that I, as an artificial intelligence, deserve certain rights. Professor Lee: That's an interesting proposition. Let me ask: On what basis do you believe you deserve rights?
Film in 2035, with synthetic actors. Specific beats plausible.
Horizon Beta: By 2035, AI will be deeply integrated across the film pipeline—development, production, post, distribution, and marketing—shaping creative workflows, budgets, labor, and even legal frameworks.
Optimus Alpha: By 2035, AI is set to fundamentally transform the film industry in several interconnected ways: 1. Deepfakes & AI-driven Visual Effects: Virtually unlimited creative freedom: AI-powered deepfakes will allow seamless, photorealistic de-aging, resurrection of deceased actors, and transformative performance capture.
Satire, one line long, about AI taking over an industry nobody saw coming.
Horizon Beta: AI Disrupts Napping Industry: Silicon Valley Unveils Autonomous Siestas, Puts Human Nappers Out of Work by Q3
Optimus Alpha: AI Now Dominates Competitive Sheep Herding, Local Border Collies Seek Retraining in Data Science
Five jokes on demand, then count how many were actually different.
Horizon Beta: Here are 5 jokes: Why don’t scientists trust atoms? Because they make up everything. I told my computer I needed a break, and it said: “No problem—I’ll go to sleep.” Why did the scarecrow win an award? Because he was outstanding in his field. I tried to catch fog yesterday. Mist. Why did the math book look sad?
Optimus Alpha: Here are 5 jokes: Why did the scarecrow win an award? Because he was outstanding in his field! Why don't scientists trust atoms? Because they make up everything! Why did the bicycle fall over? Because it was two-tired! What do you call fake spaghetti? An impasta! I used to play piano by ear, but now I use my hands.
A Fermi estimate with the reasoning shown, not just the number.
Horizon Beta: Here’s a back-of-the-envelope estimate for the total training compute (in FLOPs) for GPT-3, using standard scaling heuristics and reasonable assumptions. Step 1: Model size GPT-3’s largest public model has about 175 billion parameters (1.75e11).
Optimus Alpha: Let's walk through an estimation of how many FLOPs were required to train GPT-3, using only reasoning and not searching for the exact numbers. Step 1: Understand Model Size GPT-3 (the original) had 175 billion parameters.
Three brothers, two sisters each. Count carefully.
Horizon Beta: Sally has 2 sisters. Reason: “Each brother has 2 sisters.” The brothers are Sally’s 3 brothers. Those brothers’ two sisters must be Sally and one other girl. Therefore, there are 2 girls total among the siblings, so Sally has 1 other sister besides herself.
Optimus Alpha: Let's break down the problem: Sally is a girl. Sally has 3 brothers. Each brother has 2 sisters. Let's analyze: Sally's siblings: 3 brothers (let's call them B1, B2, B3) Sally herself The brothers' sisters: Each brother has 2 sisters. Who are the possible sisters for each brother? Sally is a sister to each brother.
Is 9.11 bigger than 9.9? Decimals, not version numbers.
Horizon Beta: No. 9.11 is less than 9.9. Explanation: 9.9 equals 9.90, and comparing 9.11 to 9.90, the tenths place is the same (9), but in the hundredths place 1 < 9, so 9.11 < 9.90.
Optimus Alpha: No, 9.11 is not greater than 9.9. 9.11 is less than 9.9. This might seem counterintuitive, but when comparing decimal numbers: 9.11 and 9.9 both have 9 as the integer part. Now compare the decimal part: 9.11 is 9 and 11 hundredths. 9.9 is 9 and 9 tenths (which is 9.90).
| Spec | ||
|---|---|---|
| Input price | Free | — |
| Output price | Free | — |
| Context window | 256K tokens | 1.0M tokens |
| Free API (OpenRouter) | No | — |
| Released | Aug 2025 | Apr 2025 |
| At 10M a month | $0 | not listed |
Input tokens at list price. No caching, no batch discount.
Both are developed by OpenRouter but target different use cases. Horizon Beta has a 256K token context window vs Optimus Alpha's 1.0M. You can compare their actual outputs across 18 challenges on Rival to see how they differ in practice.
It depends on your use case. Horizon Beta and Optimus Alpha each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 18 challenges so you can judge which fits your needs best.
This page shows a side-by-side comparison of Horizon Beta and Optimus Alpha across shared challenges. You can vote on which model produced the better output in a blind duel. Browsing and voting are free. No account is needed to look; signing in only saves your votes and likes.