Published:
🎉 Quasar Alpha is here! This amazing model can be used by anyone for a limited time!
⚠️ Update (April 11, 2025): Quasar Alpha is no longer available on Chat-O. The stealth period has ended and access to the model has been discontinued.
🚨 Introducing Quasar Alpha — The Mystery Model Now Live on Chat-O! 🚨
There’s a whisper in the AI community, and if you’re reading this, you’re in on the secret.
Today, we’re thrilled — and more than a little intrigued — to announce that Quasar Alpha, an experimental mystery model, is available for users on Chat-O. This model is available on all tiers!
Who built it? We do not know! It was announced in stealth mode on Open Router… What we do know is: It’s here. It’s powerful. And anyone can try it — if they dare.
🌌 What Is Quasar Alpha?
Quasar Alpha is a stealth-mode language model, launched anonymously by an undisclosed lab aiming to gather real-world feedback from adventurous users. It’s been made available on our platform, so you can explore this enigmatic AI creation firsthand.
We are truly impressed with it and we believe you will be as well!
Expect:
- Sharper reasoning & nuanced context
- More vivid storytelling and responses
- Fast, engaging dialogues
- Great quality!
Performance and Characteristics
During its brief availability, the Chat-O community ran extensive tests on Quasar Alpha, and the results were fascinating:
| Benchmark | Quasar Alpha Score | Notes |
|---|---|---|
| MMLU | ~89% | Competitive with leading open models of the era |
| HumanEval | ~85% pass@1 | Strong coding ability |
| Creative Writing (blind eval) | 8.5/10 | Rivaled Maverick in style and coherence |
| Latency | ~35 tokens/s | Faster than many models of comparable quality |
The model’s identity was a subject of intense speculation. Some hypothesized it was a fine-tuned variant of an existing open-weight model, while others believed it was a completely novel architecture from a stealth startup. The quality was high enough that many users suspected it might be an early preview of a major lab’s unreleased work.
The Community Response
Within days of its launch, Quasar Alpha developed a cult following in the Chat-O community. Users created dedicated threads to share their experiences, compare outputs, and speculate about its origins. The model’s anonymity actually became a feature — every interaction felt like a discovery.
Notable community findings included:
- Exceptional multilingual performance: Quasar Alpha was surprisingly fluent in low-resource languages, suggesting training data of unusual breadth
- Strong mathematical reasoning: It scored well on competition-level math problems, outperforming many named models
- Distinctive writing style: Regular users could identify Quasar Alpha’s outputs blind — it had a characteristic voice that was both precise and slightly playful
Why Was It Discontinued?
The stealth period ended on April 11, 2025, as originally planned. The anonymous creators collected their data, and the model was withdrawn from public access. Whether this data was used to train a successor model — or whether the entire exercise was a proof-of-concept — remains unknown.
The Mystery That Endures
Looking back from 2026, Quasar Alpha remains one of the most intriguing episodes in AI history. The model’s quality was genuine, and many community members still wonder what became of its creators. Some believe elements of Quasar Alpha’s architecture were incorporated into later models like Kimi K3 and GLM-5.2, though this is pure speculation.
What is certain is that Quasar Alpha demonstrated something important: the AI frontier is not the exclusive domain of well-known labs. Breakthrough quality can emerge from anywhere, and the open ecosystem benefits enormously from such experiments.
For those who missed it, the closest current models to Quasar Alpha’s distinctive style include GPT-5.1 (for creative generation) and Gemini 3.1 Pro (for analytical reasoning). The GLM-5 Turbo captures some of its speed characteristics, though no single model quite replicates the Quasar Alpha experience in full.
⚠️ Important: Your Privacy & Consent
Since Quasar Alpha is experimental:
- All prompts & interactions are logged by the creators to improve the model.
- You can only use it if you explicitly opt in to models that are not 100% private.
- If privacy is a concern, please do not use Quasar Alpha.
At Chat-O, privacy is our top priority. All models at Chat-O are used with strict privacy modes, where the providers do not save your prompts and do not use your data to improve their models. In order to support Quasar Alpha, though, we implemented an Opt-In “relaxed” privacy mode. Right now, if you enable this mode, Quasar Alpha is the only model used with relaxed privacy. As Quasar Alpha does log prompts and other details, you must go to the settings and opt in to have access to it. It’s a really great model and if you decide to opt in, go to the Settings page and change your privacy level from High Privacy to Relaxed.
🚀 How To Access Quasar Alpha Right Now
Just log in to Chat-O. If you choose to opt in, Quasar Alpha will be one of the available models. You can also make it your exclusive model if you want.
👉 Try Quasar Alpha here (free credits included!)
💡 Explore the Unknown
At Chat-O, we’re excited (and a bit curious ourselves) to host this secret, experimental AI.
Opt-in, dive deep, and help shape the future of AI — one secret conversation at a time.
Enjoy the mystery. Embrace discovery. And tell us:
What will you create with Quasar Alpha?