Union Alpha AI: Benchmarks, Specs, Cost & How to Use It

Updated October 10, 2026. Original research snapshot: September 16, 2026.

Route and pricing update, November 2, 2026: The old Pareto URL now redirects to Pareto Latest, an alias that follows newer family models and has no fixed price of its own. It currently points to Pareto 26.10. That pinned listing advertises US$0.80 input, US$3.20 output and US$0.03 cache reads per million tokens, a 1,048,576-token context and up to 131,072 completion tokens. It supports tools but does not support response_format. The September 16 launch and October 10 Pareto observations below are historical. Do not use their old 262K window, $2.50/$7.50 rate or JSON-mode description as instructions for today’s latest alias. The alias relationship does not establish that Pareto 26.10 is the exact model served during the Union Alpha preview. No new model request or Kingy benchmark was run.

Union Alpha launched as an anonymous model in the OpenCode and OpenRouter ecosystems. OpenRouter now identifies it as Unbiased Pareto and says the free preview ended. The original listing documented text and image input, a 262,144-token context window and tool support.

That makes it worth trying, but this review does not establish that it is a frontier model.

The September 16 listing did not disclose a model card, technical report, parameter count, weights, license, knowledge cutoff or benchmark scores. OpenRouter now names Unbiased as the provider and says OpenRouter itself was not the developer or owner. The phrase “frontier-level performance” appears in the listing, but this review found no benchmark table in the October 10 Pareto listing.

For a current evaluation, use the named Pareto route with a spending limit, sanitized inputs and your own task set. Check live terms and reliability before production use. The free Union Alpha preview recommendation below describes the September 16 launch window.

Union Alpha at a glance

This table preserves September 16 launch identifiers and prices together with the October 10 Pareto snapshot. Its 262K window and structured-output descriptions do not describe Pareto 26.10 or establish the behavior of today’s latest alias.

The launch details came from the OpenRouter model page, OpenRouter’s model API metadata, and the OpenCode Go documentation. The model page and API catalog are the useful sources here because they expose the actual integration surface rather than repeating community guesses.

What Union Alpha is, and what it is not

Union Alpha launched as a hosted API model under a stealth release program. It is not currently an open-source or open-weight model.

That distinction matters. “OpenCode” and “OpenRouter” describe the tools and gateways through which you can access the model. They do not mean that Union Alpha’s weights or training materials are open. The October 10 Pareto metadata has no Hugging Face model ID, no license, and no architecture description beyond text+image->text.

OpenRouter has since identified the provider as Unbiased and the model as Pareto. That reveal resolves the earlier provider-identity uncertainty. It does not disclose the underlying architecture, parameter count, training data or weights.

At launch, Union Alpha was an anonymous hosted multimodal preview. OpenRouter now identifies it as Unbiased Pareto and says the free stealth period ended.

The published specifications, explained

A 262K context window

The October 10 Pareto listing gives a 262,144-token context window. That is large enough for a serious repository, a long technical report, a multi-file migration, or a substantial research bundle in one conversation.

In the September 16 comparison snapshot, Union Alpha’s 262,144-token window was large but not the biggest available. GPT-6 Astra, Claude Opus 4.7, DeepSeek V4, Qwen3.8, and GLM-5.3-Flash were listed with roughly one million or more tokens on hosted routes. A larger window does not guarantee that a model will retrieve details from the middle of a long prompt. Context capacity and context quality are different measurements.

Pareto’s October 10 262K window is still useful. Most coding tasks do not need a million tokens, and you can keep prompts focused without paying to fill a huge context window. The practical question is whether it can identify the right files, keep state across tool calls, and finish the task without losing important constraints. This review did not independently evaluate those context and task abilities.

Image input, text output

The October 10 Pareto listing accepts text and images and returns text. That makes it suitable for screenshot-based UI debugging, diagram interpretation, document images, visual error messages, and code tasks that include a visual reference.

Video is not listed for the current Pareto route. Do not assume that Pareto can inspect video simply because some neighboring models can.

Tools and structured output

The October 10 Pareto listing includes tools and tool_choice as supported parameters. That means the API surface can participate in an agent loop: the model can request a function, receive the result, and continue the conversation.

The October 10 listing also supported response_format for JSON output, but said that JSON-schema enforcement was not available. This is historical API metadata: the November 2 Pareto 26.10 listing does not support response_format. Your client should validate every response, retry malformed output, and treat model-generated JSON as untrusted input.

Tool support is a protocol feature, not a quality score. A model can accept a tool schema and still choose the wrong tool, produce invalid arguments, stop early, or loop. Tool-call success rate, task completion, and recovery from errors are the metrics to watch in a real agent.

No public reasoning knob

The October 10 Pareto metadata does not list reasoning, reasoning_effort, or a similar control. That establishes only the public API surface; it does not establish how the model reasons internally.

The absence of a documented control makes Pareto harder to compare with models that expose low, high, and maximum reasoning effort. It also means you may have less control over latency and token use. A short coding fix and a difficult multi-step investigation may receive the same hidden inference policy.

Missing technical fields

The October 10 Pareto listing leaves several fields blank or generic:

• no parameter count or active-parameter count;
• no dense-versus-mixture-of-experts description;
• no tokenizer identity;
• no training-data or training-recipe description;
• no knowledge cutoff;
• no public model card or license;
• no weights or local deployment path.

Those omissions are not proof that the model is weak. They are proof that outsiders cannot yet audit or reproduce it.

Benchmarks and evaluations: the honest answer

The October 10 Pareto listing and model metadata do not publish a benchmark table or expose Artificial Analysis or Design Arena scores for the model.

That leaves us with a description, a specification, and live service metadata. The description says Union Alpha is intended for research, coding, agentic workflows, and broad general-purpose work. Those are intended use cases, not results.

The difference is easy to lose in launch coverage:

OpenRouter’s public benchmark system separates benchmark scores from token usage and exposes provider-level signals. Its documentation says benchmark runs can be pinned to a provider endpoint, while tool-calling, throughput, and benchmark data remain separate routing signals. For Union Alpha, treat capability, provider quality, and agent reliability as three different questions.

There is one operational detail worth preserving from the September 16 launch snapshot. The Union Alpha page then exposed provider metrics such as latency, throughput, uptime, and availability. Those launch-era values were a dashboard snapshot, not an SLA or model benchmark. For current route details, check the Pareto listing before relying on them.

Do not turn a screenshot of one successful coding task into a benchmark. A credible comparison should publish the task set, harness, model ID, prompt, tool definitions, temperature, timeout, number of trials, failures, and raw results. Even then, one harness measures one slice of behavior.

The OpenRouter benchmark API documentation and Auto Exacto documentation explain how the platform presents these signals. The provider pages for established models show the contrast: they include public index scores and, in several cases, model-card benchmarks. The historical Union Alpha listing has no comparable benchmark table; this review has not independently evaluated the current Pareto route.

Union Alpha versus frontier and open-weight models

The table below is a dated snapshot of public catalog information. The prices use the base rates returned by OpenRouter’s model catalog API on September 16, 2026, shown per 1M input/output tokens. Variant, batch, and provider-specific rates can differ and change. Artificial Analysis scores are shown only where OpenRouter exposes them, and they are not a claim about Union Alpha.

The source material behind the open-weight rows is available in the GLM-5.3-Flash model card, DeepSeek V4 Pro model card, and Qwen3.8 model card. The table uses “open weights” rather than “fully open source” because published weights do not automatically mean that the training data, training code, or complete serving stack is open.

What the comparison means in practice

Against frontier models: Union Alpha was free during the September 16 preview and offered enough context for many coding sessions. That price advantage has ended. Compare the current Pareto rate, task performance and provider reliability with named frontier routes before choosing one.

Against open-weight models: Union Alpha is easier to try because someone else runs the hardware. Qwen3.8, DeepSeek V4, and GLM-5.3-Flash give you weights, licenses, model cards, and a path to self-hosting or choosing among multiple inference providers. Self-hosting is not free, though: memory, GPUs, electricity, engineering time, and maintenance become your problem.

Against low-cost hosted open models: GLM-5.3-Flash and Qwen3.8 offered an economic alternative during the anonymous preview. The free window has since closed; compare their live prices with the named Pareto route before buying. Those models publish weights and documentation that can help evaluation.

On multimodality: Union Alpha handles images. Qwen3.8 and GLM-5.3-Flash publish broader visual capabilities, including video input in their model cards or catalog descriptions. A model supporting more media types is not automatically better at the media you care about, but the distinction matters for product design.

On long context in the historical comparison: Union Alpha’s September 16 listing and the October 10 Pareto snapshot gave 262K. The November 2 Pareto 26.10 listing gives 1,048,576 tokens. A listed capacity does not establish retrieval quality or reliable performance on a particular repository.

The Ox Alpha precedent does not identify Union Alpha

Ox Alpha previously showed that a stealth listing could later be attributed to a named provider. OpenRouter has now made Union Alpha’s own attribution: Unbiased Pareto. The Ox example is historical context, not evidence that the two models are related.

The documented profiles differ. GLM-5.3-Flash is listed with a 1.3-million-token context window, text/image/video input, a 320B-parameter total model with 18B active parameters, and published MIT-licensed weights.

Union Alpha is listed with 262K context, text/image input, no video, no disclosed parameter count, and no weights. Those differences could reflect a different model, a different service configuration, or a deliberately reduced preview surface. They cannot establish identity.

The useful conclusion is narrower: stealth previews can be real model launches, but their names are not evidence. OpenRouter has now made the provider claim: Union Alpha was Unbiased Pareto. Independent behavior and capability comparisons still require reproducible tests.

How Union Alpha was accessed at launch

The following OpenCode and OpenRouter examples document September 16 routes. Check the current Pareto listing and your account before sending a paid request; do not assume the old free route or model ID still works.

OpenCode Go

At launch, OpenCode Go listed Union Alpha Free. Go remains a separate $10/month subscription, but its current public model list does not include Union Alpha. OpenCode Zen’s current public model list also omits it; neither list establishes an active Union Alpha entitlement for a particular account. The terminal and API examples below are historical and should not be used as current availability instructions.

In the OpenCode terminal UI:

opencode
/connect        # choose OpenCode Go and paste your Go key
/models         # choose Union Alpha Free

At launch, the Go API documentation gave this historical route:

Model ID:  opencode-go/union-alpha
Endpoint:  https://opencode.ai/zen/go/v1/messages
Protocol:  Anthropic Messages API

That protocol detail matters. The historical OpenCode Go route used a different model ID and request shape from the OpenRouter OpenAI-compatible stealth route. OpenCode also asks compatible clients to send a stable x-opencode-session value so it can optimize prompt caching and routing.

OpenCode Go is not a free subscription. The Union Alpha model usage was listed as free for a limited time; the Go plan remains a separate $10/month product. If you use Zen’s pay-as-you-go route instead, check the current Zen setup and pricing page, which currently advertises a $20 balance purchase plus a card-processing fee.

OpenRouter

At launch, OpenRouter’s stealth model ID was:

stealth/union-alpha

The API is OpenAI-compatible:

curl \
  https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stealth/union-alpha",
    "messages": [
      {
        "role": "user",
        "content": [
          {"type": "text", "text": "Describe the main issue in this screenshot."},
          {"type": "image_url", "image_url": {"url": "https://example.com/screenshot.png"}}
        ]
      }
    ]
  }'

OpenRouter says the Union Alpha free stealth period has ended. The October 10 Pareto snapshot lists $2.50 per million input tokens and $7.50 per million output tokens; check the live listing before use. Historical free-model limits no longer establish a current Union Alpha allowance. OpenRouter’s current rate-limit documentation says free-model usage limits depend on account and credit status and can include platform-level caps.

For new requests, check the named Pareto listing, its current account limits and spend controls. The old free-model allowance and stealth removal terms describe the launch period.

OpenRouter also offers a chat playground, which is the quickest way to test image understanding and basic coding behavior without writing an integration. For an application, set a spending limit anyway, record the returned model and provider metadata, and handle rate limits and empty responses.

Cloudflare AI Gateway

The Cloudflare route documented for the launch example is Cloudflare AI Gateway proxying the OpenRouter provider; current Pareto use needs a fresh model-ID and account check. It is not evidence that Union Alpha is hosted natively inside Cloudflare Workers AI.

Cloudflare documents the OpenRouter endpoint as:

https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/openrouter

A compatible request looks like this:

curl \
  "https://gateway.ai.cloudflare.com/v1/$CLOUDFLARE_ACCOUNT_ID/$CLOUDFLARE_GATEWAY_ID/openrouter/v1/chat/completions" \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stealth/union-alpha",
    "messages": [
      {"role": "user", "content": "Review this migration plan for missing steps."}
    ]
  }'

Cloudflare AI Gateway can add logging, analytics, caching, rate limiting, retries, and other gateway controls. Those controls are useful for an application, but they do not change the upstream model’s identity, benchmark record, or terms. The Cloudflare OpenRouter integration guide requires an active OpenRouter token and the OpenRouter model name.

Do not infer a native Cloudflare Workers AI model ID from an OpenRouter listing. The historical Cloudflare AI Gateway pattern used OpenRouter as the upstream. Check Cloudflare’s current integration guide and Pareto’s live model ID before configuring a route.

Privacy: “not used for training” is not the same as zero exposure

Privacy and data retention depend on the route.

OpenRouter’s general data-collection documentation says OpenRouter does not store prompts or responses unless a user opts into logging or product improvement. That statement applies to OpenRouter’s own systems; the upstream provider can have a separate policy. Its provider-logging documentation makes the same separation.

The Stealth Program EULA deserves a close read. It says user content may be collected by OpenRouter and shared with a stealth provider, and that stealth models are temporary. Its acceptable-use policy also prohibits sending sensitive information such as health or financial data, bypassing rate limits, and publicly disseminating confidential technical information about model performance.

For a hosted model, use a disposable branch or a sanitized repository. Remove API keys, credentials, customer data, personal data, private prompts, unreleased product details, and regulated information. “The provider says it does not train on data” does not mean “the data never leaves your machine.”

How to evaluate Union Alpha properly

If you want to know whether Union Alpha is good, test the work you actually need it to do. A single clever prompt cannot answer that question.

Use a clean copy of a repository and a fixed task set. Include at least:

1. repository orientation: ask the agent to locate the right files and explain the relevant code paths;
2. a small bug fix with an existing test;
3. a multi-file change with a clear acceptance test;
4. a refactor that must preserve behavior;
5. a task requiring a tool call with structured arguments;
6. a task that deliberately returns a tool error and tests recovery;
7. screenshot-to-code or visual UI debugging;
8. long-context retrieval from several documents or modules;
9. a factual question with a known answer and a verification source;
10. an ambiguity test where the correct behavior is to ask for clarification;
11. a task that should be refused or escalated because it touches a secret;
12. a final review in which the model must inspect its own diff and tests.

Record:

• time to first token and total wall-clock time;
• task completion, not merely a plausible explanation;
• tests passed and tests added;
• number of tool-call errors and retries;
• human interventions and prompts needed to recover;
• files changed and unwanted changes in the diff;
• context length and whether important details were retrieved;
• factual errors, unsupported claims, and citation failures;
• service errors, rate limits, truncations, and empty responses.

Run each important task more than once. Compare Union Alpha with one frontier model and one open-weight model through the same agent harness. Keep the prompt, tools, timeout, temperature, repository state, and acceptance criteria fixed. Evaluate the model-plus-provider-plus-harness combination, because changing the harness can change the result as much as changing the model.

You can use the results to choose a tool for yourself. Before publishing detailed performance data, review the Stealth EULA and make sure the information is public or you have permission to disclose it.

Who should use Union Alpha?

Union Alpha is a good fit for:

• developers evaluating the named Pareto route with a defined spending limit;
• visual debugging and screenshot-based implementation work;
• side projects and disposable branches;
• long prompts that fit comfortably inside 262K tokens;
• private evaluations of coding-agent behavior;
• research and drafting where a human checks the facts.

It is a poor fit for:

• production systems that need a stable model name and published SLA;
• sensitive repositories or regulated data;
• unattended deployment without a human review gate;
• applications that require strict JSON schemas;
• teams that need local inference or a clear license;
• budgets that cannot absorb the current paid Pareto rate or changing provider terms.

Frequently asked questions

What is Union Alpha?

Union Alpha was a multimodal stealth listing on OpenRouter and OpenCode Go. OpenRouter now identifies it as Unbiased Pareto. The former Go listing is absent from OpenCode’s current public model list.

Who made Union Alpha?

OpenRouter says Union Alpha was developed and operated by Unbiased and revealed as Pareto. OpenRouter routed requests and was not the developer, owner or provider.

Is Union Alpha free?

No current free Union Alpha offer is established by these sources. OpenRouter says the free stealth period ended and lists the named Pareto route at $2.50/M input, $7.50/M output and $0.25/M cache read as of October 10, 2026. OpenCode Go is a separate $10/month subscription; its current public model list omits Union Alpha.

How long will the free preview last?

The launch notice described a limited free period. OpenRouter now says it ended. The original announcement did not give a fixed end date, so this review does not invent one.

What is the context window?

The October 10 Pareto listing gives 262,144 tokens, with up to 131,072 completion tokens.

Does it support images and video?

The October 10 Pareto listing supports text and image input and text output. Video input is not listed.

Does it support tool calling?

Yes. The October 10 Pareto listing includes tools and tool_choice. That confirms API compatibility, not reliable task completion.

Does it support JSON mode?

The October 10 Pareto listing accepts response_format, but says JSON-schema enforcement is not available. Validate every response yourself.

Is Union Alpha open source or open weight?

Not according to the October 10 Pareto listing. No weights, license, model card, or local deployment package is linked there.

Is it available through Cloudflare AI?

Cloudflare AI Gateway documents an OpenRouter integration, but the old stealth/union-alpha example is historical. Check the current latest alias or select the pinned unbiased/pareto-26.10 route and Cloudflare’s supported provider configuration. No native Workers AI @cf model ID was verified here.

Is it safe to use with private code?

Treat it as unsafe for sensitive material until you have reviewed the exact route and terms. At launch, OpenCode documented zero retention for its Go route. OpenRouter’s Union Alpha listing says the upstream provider may retain prompts and completions, even though it will not use them for training. Current Pareto route terms must be checked separately. The historical Stealth EULA restricted sensitive inputs; review the current Pareto route terms separately.

Is Union Alpha better than GPT-6 Astra, Claude, DeepSeek, Qwen, or GLM?

The October 10 Pareto listing publishes no benchmark table that supports that conclusion. It may be useful, fast, or surprisingly capable, but this review has no independent Kingy benchmark for it.

Bottom line

Union Alpha was a free look at an unnamed model with a large context window, image input and agent-facing tools. The provider is now named and the free period has ended. Its capability remains unverified by an independent Kingy evaluation.

Try it on work you can throw away. Save the diffs, measure task completion, verify the facts, and keep the provider’s uncertainty in view. For the paid Pareto route, compare repeatable task results, provider reliability and actual cost against alternatives before relying on it.

Sources and research notes

Research snapshot: September 16, 2026. Prices, provider availability, and preview access can change quickly.

• OpenRouter: Union Alpha model page
• OpenRouter: model catalog API
• OpenCode: Go documentation
• OpenCode Zen
• Cloudflare: OpenRouter through AI Gateway
• Cloudflare: OpenCode integration
• OpenRouter: Stealth Program EULA
• OpenRouter: API credit and rate limits
• OpenRouter: data collection
• OpenRouter: provider logging
• OpenRouter: benchmark API
• OpenRouter: Auto Exacto benchmark and routing methodology
• GPT-6 Astra on OpenRouter
• Claude Opus 4.7 on OpenRouter
• GLM-5.3-Flash on OpenRouter
• GLM-5.3-Flash model card
• DeepSeek V4 Pro 0813 on OpenRouter
• DeepSeek V4 Pro 0813 model card
• Qwen3.8 27B on OpenRouter
• Qwen3.8 27B model card