Union Alpha API: 262K Multimodal Model

A multimodal model for research and agents

Union Alpha is a multimodal model available on TokenRa as union-alpha. It accepts text and image input, returns text, and provides a 262,144-token context window with up to 131,072 tokens of output. The provider is undisclosed and listed as Stealth.

Union Alpha specifications

The values below reflect the specifications published for this model. Confirm the live limits and availability for your account in the TokenRa console before relying on them.

• Model ID: union-alpha; provider listed as Stealth and operated independently of TokenRa.
• Released September 16, 2026; billed at $5 per 1M input tokens and $14 per 1M output tokens on TokenRa.
• Context window 262,144 tokens; maximum output 131,072 tokens.
• Input and output: text plus image in, text out.
• Tool calling supported through tools and tool_choice; structured output through response_format, with JSON schema accepted but not strictly enforced.
• Reported performance: around 17.6 seconds P50 latency, near 14 tokens per second throughput, and about 98% uptime over a three-day window.

Where Union Alpha fits

Union Alpha favors depth and context length over response speed, which shapes where it works well.

• Batch processing and asynchronous jobs where latency is not user-facing.
• Long-form generation that benefits from a large output ceiling.
• Repository-scale code reading and analysis within a single context.
• Agentic workflows with tool calls, clear stopping conditions, and application-level validation.
• Exploratory research and prototyping where context length and output size matter more than speed.

How to get a Union Alpha API key

Register for TokenRa, open the dashboard, create an API key, confirm that Union Alpha is enabled for your account, and then send an OpenAI-compatible request from your server. Keep the key private and verify the live model identifier before production use.

curl https://tokenra.io/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"union-alpha","messages":[{"role":"user","content":"YOUR_TEST_PROMPT"}]}'

Any OpenAI-compatible client works the same way: point the API base at https://tokenra.io/v1 and set the model ID to union-alpha.

import openai

client = openai.OpenAI(
    base_url="https://tokenra.io/v1",
    api_key="YOUR_API_KEY",
)
response = client.chat.completions.create(
    model="union-alpha",
    messages=[{"role": "user", "content": "YOUR_TEST_PROMPT"}],
)
print(response.choices[0].message.content)

Union Alpha pricing: $5 input, $14 output

Union Alpha is billed per token through TokenRa: $5 per 1M input tokens and $14 per 1M output tokens. Your account needs a positive balance before it can send requests: the minimum top-up is $10, and credits do not expire while the account is active. Access is also subject to account eligibility, provider availability, and applicable rate limits. Verify the live TokenRa console before production use.

What to watch out for

Review these points before depending on Union Alpha in production.

• High latency: around 17.6 seconds at P50, which rules out real-time chat and low-latency agents.
• Modest throughput: near 14 tokens per second, so streaming responsiveness will feel slow.
• Availability: about 98% uptime over a three-day window, measured shortly after launch.
• Undisclosed provider: no published company, model card, or research paper to evaluate.
• Structured output is not schema-enforced, so validate and repair JSON in your application.
• Pricing can change; design for a fallback model.

Production checklist

• Keep API keys on the server and separate credentials by environment.
• Confirm the 262K-token context and supported input modalities against the live integration before relying on them.
• Set generous timeouts that account for the reported P50 latency, plus bounded retries and tracing.
• Validate structured output against your own schema rather than trusting response_format.
• Configure a fallback model for availability gaps and for any change to pricing.
• Review data handling, provider terms, privacy requirements, and output rights for your use case.

Union Alpha API FAQ

Continue exploring