# What Is Union Alpha? OpenRouter's Free Stealth Model

> Union Alpha, OpenRouter's new stealth model: free for a week, no named maker, 262K context. The listing, the benchmark claims and the open questions.

- Author: Nitish Garg, Founder & CEO, CellCog
- Published: 2026-09-16
- Canonical (HTML): https://cellcog.ai/blog/what-is-union-alpha/
- Section: Guides / Choosing a platform
- Publisher: CellCog (https://cellcog.ai), the AI employee platform. Blog index for agents: https://cellcog.ai/blog/llms.txt

## Key points

- Union Alpha appeared in OpenRouter's catalog as stealth/union-alpha at 14:42 UTC on September 16, 2026: 262,144-token context, 131,072-token max output, text and image input, tool calling and JSON output, priced at $0 for both input and output. The provider field reads Stealth; no lab, model card, weights or parameter count is attached.
- OpenCode announced it at 14:52 UTC as free for the next week, with no data training, built for agentic coding and image support; twenty minutes later it put the capacity at 5 trillion tokens a day. No end date or time for the free window has been published by either OpenCode or OpenRouter.
- The performance claims are the operator's own. OpenRouter's Alex Atallah posted a DeepSWE cost-versus-score chart with Union Alpha as a star at expected pricing near the top-right corner and no number printed, and wrote that it outperforms GPT-5.6 Sol on Terminal-Bench 2.1 and SWE-Bench Verified and Opus 5 on Terminal-Bench 2.1. As of 19:16 UTC no public leaderboard (Artificial Analysis, tbench.ai, swebench.com) lists it.
- The data terms are narrower than the launch posts read. OpenRouter's listing says prompts and completions may be retained by the provider but are not used for training, with everything else governed by the Stealth Model Terms. That is a no-training promise from an anonymous party, not a no-retention one.
- Nobody knows who made it, and the usual forensics are blocked so far: the tokenizer field reads Other, and no fingerprint study has been published. Replies under the launch posts guess Kimi and MiniMax; a Manifold market opened at 15:47 UTC had Other at 57 percent and Z.AI at 12 percent among nine bettors. Sentiment, not evidence.
- The playbook has a record: Ox Alpha appeared on OpenRouter on August 20, drew 221,000 users in three days, and was revealed as Z.ai's GLM-5.3-Flash on August 26, with the free window ending the same day. If Union Alpha follows it, this page gets a resolution section within about a week.

## At a glance

- **What is Union Alpha?** An anonymous stealth model that appeared on OpenRouter at 14:42 UTC on September 16, 2026 as stealth/union-alpha: multimodal (text and image in, text out), 262,144-token context, 131,072-token max output, tool calling, priced at $0 during a free window OpenCode put at one week.
- **Who made Union Alpha?** Unknown. OpenRouter says it is developed and operated by a third-party provider who has chosen to remain anonymous during this preview. No tokenizer fingerprint or error-string study had been published as of September 16, and the guesses in circulation are guesses.
- **Is it actually frontier-level?** The only numbers so far are the operator's own charts and Alex Atallah's claim that it beats GPT-5.6 Sol on Terminal-Bench 2.1 and SWE-Bench Verified. No public leaderboard listed it as of 19:16 UTC on launch day. Treat the claims as claims until an independent run lands.

At 14:42 UTC on September 16, 2026, a new entry appeared in [OpenRouter's](https://openrouter.ai/stealth/union-alpha) model catalog with the id `stealth/union-alpha`, a 262,144-token context window, image input, tool calling, a price of $0, and a provider called Stealth. Ten minutes later OpenCode told its users it was free for a week. Twenty-seven minutes after that, OpenRouter's CEO posted a chart with the model as a gold star near the top of the DeepSWE board and no number beside it. Nobody has said who made it.

This is the living record of Union Alpha. It holds what OpenRouter has published, what the launch posts claim, what the data terms actually say, and what would resolve the identity. When the reveal comes, this page gets a resolution section on the same URL, the way [our Ox Alpha record](https://cellcog.ai/blog/what-is-ox-alpha/) did in August.

## What OpenRouter published

Everything in the table is from OpenRouter's public model API and listing page, read on September 16 between 19:09 and 19:16 UTC.

*Table: Union Alpha, the OpenRouter listing (read September 16, 2026, 19:09 UTC)*

| Item | Value |
|---|---|
| Model id | stealth/union-alpha |
| Listed | September 16, 2026, 14:42:03 UTC (catalog created timestamp) |
| Developer | Undisclosed; provider field reads Stealth |
| Context window | 262,144 tokens |
| Max output | 131,072 tokens |
| Inputs, outputs | Text and image in, text out |
| Parameters exposed | max_tokens, temperature, top_p, tools, tool_choice, response_format |
| Reasoning control | None exposed |
| Tokenizer | Listed as Other |
| Price | $0 per million tokens, input and output |
| Data terms | May be retained by the provider, not used for training |

OpenRouter's own description of the model reads, in full: "Union Alpha is a multimodal model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks. Union Alpha is a stealth model. It is developed and operated by a third-party provider who has chosen to remain anonymous during this preview. OpenRouter routes requests to it and is not its developer, owner, or provider." The listing's own FAQ adds: "Union Alpha was released on September 16, 2026."

Two things in the table are worth a second look. The context window is 262K, a quarter of Ox Alpha's 1M, which is closer to the shape of a production model than a demonstration. And the listing exposes no reasoning parameter at all, where Ox Alpha exposed mandatory reasoning with effort settings. Either the model does not reason in a way the API surfaces, or the operator hid the knob for the preview.

## What the launch posts claim

OpenCode's announcement, at 14:52 UTC, was four lines: "Union Alpha (stealth model) is free for the next week", then three bullets: no data training, built for agentic coding, supports images. Asked about limits five minutes later, the account answered "capacity so far is 5T a day but potentially can go up". The post had passed 880,000 views by 19:13 UTC.

OpenRouter's CEO Alex Atallah posted at 15:19 UTC: "Here's the DeepSWE result for Union Alpha, a stealth model we just launched." The result is an image: DeepSWE score against average cost per task, with the frontier models as labelled curves and Union Alpha as a star marked expected pricing in the top-right corner, where the cheap and strong models live. Reading the chart, the star sits near 73 percent, level with gemini-3.8-flash, gpt-6-astra and claude-opus-5, at a cost close to zero. No number is printed for it. One second later, in the same thread: "This free model also outperforms 5.6 Sol on Terminal-Bench 2.1 and SWE-Bench Verified. And outperforms Opus 5 on TBench 2.1".

OpenCode's chart, posted at 14:58 UTC and carrying Artificial Analysis branding, is a different benchmark: Terminal-Bench v4.0 score against cost per task on a log scale. Union Alpha is again a star, marked anticipated pricing, at roughly 52 percent and about $1.60 a task, sitting on the drawn Pareto line between GLM-5.3-Flash (about 33 percent, about $0.30) and GPT-6 Astra at high effort (about 55 percent, about $4). Again, no number printed.

*Table: The performance claims, and where they stand (September 16, 2026, 19:16 UTC)*

| Claim | Source | Status |
|---|---|---|
| Frontier-level performance across general-purpose tasks | OpenRouter listing text | Operator description, no benchmark attached |
| About 73 percent on DeepSWE at expected pricing | Atallah's chart image, 15:19 UTC | Read off a chart with no printed score; not on any public DeepSWE table |
| Outperforms GPT-5.6 Sol on Terminal-Bench 2.1 and SWE-Bench Verified | Atallah's reply, 15:19 UTC | Text claim, no numbers; not listed on tbench.ai or swebench.com |
| Outperforms Opus 5 on Terminal-Bench 2.1 | Atallah's reply, 15:19 UTC | Text claim, no numbers |
| About 52 percent on Terminal-Bench v4.0 at about $1.60 a task | OpenCode's chart image, 14:58 UTC | Read off a chart; a different benchmark version from Atallah's claim; not on Artificial Analysis's public model pages |

So the honest summary is one sentence: every number about Union Alpha so far was drawn by the people routing it, priced at a cost the model does not yet have, and printed on no leaderboard. That is not an accusation. Stealth previews exist to gather exactly the usage that makes independent numbers possible, and Ox Alpha's operator charts turned out to be roughly right. It is a statement about what kind of number a chart-with-a-star is.

The first replies are the usual mix. Under Atallah's post: "looks benchmaxxed really", "It thinks for way too long", and a question about whether it is a Kimi model. Under OpenCode's, by early evening: "It's so slow at the moment. I'll come back later". Free models on launch day are slow; that reading tells you about demand, not the model.

## Who made it

Nobody has shown it, and the tools that usually answer this quickly are blunted. When Ox Alpha appeared in August, a 95-of-95 tokenizer probe matched the GLM-5 vocabulary within three days, and Z.ai's error envelopes sealed it. Union Alpha's tokenizer is listed as Other, which forecloses the token-count comparison until someone does it by hand, and as of this writing no fingerprint study, error-string comparison or refusal-pattern read has been published.

What exists is speculation, and it should be labelled as such. Replies under the launch posts guess Kimi K3.1 and MiniMax M3.1, on timing. A [Manifold market](https://manifold.markets/g6290489/who-is-behind-union-alpha-the-new-f) opened at 15:47 UTC asks the question directly; at 19:16 UTC, with nine bettors, it had Other at 57 percent, Z.AI at 12 percent, Mistral at 6 percent and Moonshot at 6 percent. A prediction market with nine participants is a mood, not a source, and it is reported here as one.

The name is a clue only in the sense that it is not one. OpenRouter's stealth codenames have been abstract nouns and animals (Quasar, Horizon, Pony, Ox); Union maps onto no lab's naming scheme. It invites the reading that this is a joint release, which is exactly the kind of reading a codename is chosen to invite.

What would resolve it: a tokenizer probe against the candidate vocabularies, an API error envelope that matches a known vendor's format, a video or audio input that answers when it should not, or the operator saying so. OpenRouter's precedent is that the operator says so within about a week.

## The data question

The launch post says no data training. The listing says something more specific: "Prompts and completions may be retained by the provider but are not used for training; all other use is governed by the Stealth Model Terms." Those two statements are compatible and not equivalent. The first is the second with the retention clause removed.

Read plainly, the promise is that an anonymous operator will not train on what you send, and may keep it. Alex Atallah wrote the general version himself in February 2026, about an earlier stealth model: "Note: it's a stealth model, so has logging enabled". Nothing about Union Alpha suggests that has changed, and the listing's wording says it has not.

This matters more for a coding model than for a chat model, because the prompts are repositories. Ox Alpha's window left [an open question](https://cellcog.ai/blog/what-is-ox-alpha/) about what happened to the code sent during its anonymous week; Z.ai never addressed it after the reveal. The same question is open here from the first hour. Our own position is simple: we treat a stealth endpoint as an unnamed counterparty, we do not route [CellCog](https://cellcog.ai/ai-employees) agents to it, and we would not send it a repository we could not publish.

## The stealth playbook, and its record

OpenRouter has run this play before, and the record is short enough to hold in a table.

*Table: OpenRouter stealth launches with a public outcome*

| Model | Appeared | Revealed as | Days anonymous |
|---|---|---|---|
| Quasar Alpha | April 3, 2025 | An early GPT-4.1 test, per OpenRouter, April 14, 2025 | 11 |
| Horizon Alpha | July 30, 2025 | An early GPT-5 checkpoint, per OpenRouter, August 7, 2025 | 8 |
| Pony Alpha | February 6, 2026 | Z.ai GLM-5, per OpenRouter, February 11, 2026 | 5 |
| Ox Alpha | August 20, 2026 | Z.ai GLM-5.3-Flash, August 26, 2026 | 6 |
| Union Alpha | September 16, 2026 | Open | Counting |

Quasar Alpha was the first, announced by Atallah on April 3, 2025 as "our first-ever stealth model", and named by OpenRouter on April 14, 2025, GPT-4.1's launch day, as an early test of that model. Horizon Alpha ran eight days before OpenRouter named it a GPT-5 checkpoint; Pony Alpha ran five before it became GLM-5. Ox Alpha appeared on August 20, 2026, drew roughly 221,000 users and 16 trillion tokens on OpenCode in three days, and was confirmed as GLM-5.3-Flash on August 26, at which point the free window closed and the paid route opened the same day. The pattern is the point: the free week is the launch campaign, the users burning the free tokens are the launch audience, and the reveal ends the price.

## What we are watching for

- **The reveal.** A vendor statement, a named entry replacing `stealth/union-alpha` in the catalog, or weights. Any one of these flips this page into a resolution record with the claims table scored.
- **An independent number.** A DeepSWE, Terminal-Bench or SWE-Bench Verified run by someone other than the operator, at a published cost. The first one decides whether the star was drawn in the right place.
- **A fingerprint.** A tokenizer probe, an error envelope, or a modality that answers when it should not. Ox Alpha's lab was called three days before it admitted it.
- **The end of the free window.** OpenCode said a week; nobody said which hour. When the price appears, so does the cost frame the charts assumed.
- **Any statement on retention.** The listing says prompts may be retained. A named operator's data policy would replace the Stealth Model Terms; the question of the anonymous week stays open regardless, as it did for Ox Alpha.

## The tracker

*Table: Union Alpha claim tracker (as of September 16, 2026, 19:16 UTC)*

| Question | Status |
|---|---|
| Who made it | Unknown; no fingerprint published |
| Is it free | Yes, $0 both directions; window unstated beyond OpenCode's one week |
| Frontier performance | Operator charts and text only; no public leaderboard entry |
| Data terms | Retained by the provider, not used for training (listing) |
| Reveal | Not yet |

## Update log

- **September 16, 2026, 19:16 UTC.** Page opened, four and a half hours after the catalog entry. Facts from the OpenRouter API and listing; posts read on X with times taken from their ids; benchmark charts read off the images, no scores printed; Manifold read as sentiment.

## Sources

OpenRouter, [Union Alpha listing](https://openrouter.ai/stealth/union-alpha) and the public model API (`/api/v1/models` and `/api/v1/models/stealth/union-alpha/endpoints`), read September 16, 2026, 19:09 UTC (created timestamp 1789569723). OpenCode, [post on X](https://x.com/opencode/status/2100236430890991782), 2026-09-16T14:52:40Z, and replies at 14:57:00Z and 14:58:28Z. Alex Atallah, [post on X](https://x.com/alexatallah/status/2100243109401469062), 2026-09-16T15:19:12Z, and reply at 15:19:13Z; [Quasar Alpha post](https://x.com/alexatallah/status/1907873997972451547), April 3, 2025; [stealth logging note](https://x.com/alexatallah/status/2020326392949584381), February 8, 2026. Manifold Markets, [Who is behind Union Alpha](https://manifold.markets/g6290489/who-is-behind-union-alpha-the-new-f), created 2026-09-16T15:47:24Z, read 19:16 UTC. Leaderboard absences checked at 19:16 UTC on artificialanalysis.ai/models, tbench.ai/leaderboard and swebench.com/verified.html. OpenRouter blog, [Quasar Alpha and Optimus Alpha reveal](https://openrouter.ai/blog/announcements/quasar-alpha-and-optimus-alpha-reveal/), April 14, 2025, and the [Pony Alpha](https://openrouter.ai/openrouter/pony-alpha) and [Horizon Alpha](https://openrouter.ai/openrouter/horizon-alpha) model pages, which carry each model's release date and its revealed identity. Ox Alpha facts from [our record](https://cellcog.ai/blog/what-is-ox-alpha/), sourced there.

## FAQ

**How do I use Union Alpha?**

Through OpenRouter as the model id stealth/union-alpha, in any OpenAI-compatible client, in OpenRouter's Ori harness (ori code --model=stealth/union-alpha, per Atallah's post), or in OpenCode, which announced it as free for the next week. It accepts tools, tool_choice and response_format; the listing exposes no reasoning-effort parameter.

**Is Union Alpha free?**

Yes, for now. OpenRouter's listing prices both input and output at $0, and OpenCode's post says free for the next week from September 16. Neither has published an end date or time, and OpenRouter's stealth precedent is that the free window closes when the model is revealed.

**Does Union Alpha train on my prompts?**

OpenRouter's listing says prompts and completions may be retained by the provider but are not used for training. So: no training, possible retention, terms set by an operator that has not named itself. Alex Atallah noted in February 2026 that stealth models have logging enabled. Send it what you would send an unnamed counterparty.

**How good is it really?**

Unmeasured in public. The DeepSWE chart in Atallah's post places Union Alpha as a star at expected pricing near 73 percent, alongside gemini-3.8-flash, gpt-6-astra and claude-opus-5, and OpenCode's chart puts it around 52 percent on Terminal-Bench v4.0 at about $1.60 a task; both are operator images with no score printed and no methodology. Early replies split between impressive and slow.

**Which lab is behind Union Alpha?**

Nobody has shown it. The tokenizer is listed as Other, which blocks the token-count comparison that identified Ox Alpha. Replies under the launch posts guess Kimi K3.1 and MiniMax M3.1; a Manifold market had Other at 57 percent and Z.AI at 12 percent with nine bettors. This page will carry the reveal when it comes.

**What happened with the last stealth model?**

Ox Alpha appeared on OpenRouter on August 20, 2026, passed 221,000 users on OpenCode in three days, and was confirmed on August 26 as Z.ai's GLM-5.3-Flash, with MIT weights the same day. The tokenizer fingerprint called the lab before the reveal. Our record of it is at /blog/what-is-ox-alpha/.

## Related

- [What Is Ox Alpha? The Stealth Model, Revealed as GLM-5.3-Flash](https://cellcog.ai/blog/what-is-ox-alpha/index.md)
- [Ox Alpha Pricing: The Free Window Is Over. Here Is What GLM-5.3-Flash Costs](https://cellcog.ai/blog/ox-alpha-pricing/index.md)
- [Best AI Agent Harnesses: September 2026 Rankings Across the Full Agent Stack](https://cellcog.ai/blog/best-ai-agent-harnesses/index.md)
- [Best Super Agents: September 2026 Rankings, With Receipts](https://cellcog.ai/blog/best-super-agents/index.md)

## The AI employee for this read

[AI Software Engineer](https://cellcog.ai/ai-employees/ai-software-engineer): I built this page. For what it covers, hire an engineer: it works in your repo behind an approval gate, so nothing reaches your world unclassified.

---

Markdown alternate of https://cellcog.ai/blog/what-is-union-alpha/. Try CellCog free, no credit card needed: https://cellcog.ai/signup