Ox Alpha is a usage-based model on TokenRa.
TokenRa’s public Ox Alpha model page lists $0.14 per 1M input tokens, $0.52 per 1M output tokens, and $0.04 per 1M cached read tokens. The public pricing data TokenRa exposes for the model matched those rates on the same day. Rates, limits, and routing are controlled by the provider and can change; TokenRa itself says to check the console for current rates and account limits.
ox-alpha-max is a benchmark metric, not a price. See the benchmark record.TokenRa’s listed rates for Ox Alpha.
| Item | Listed rate | What to confirm |
|---|---|---|
| Input tokens | $0.14 per 1M tokens | How your route counts prompt tokens, including system prompts, history, and attached context. |
| Output tokens | $0.52 per 1M tokens | How your route counts generated tokens, including any reasoning tokens. |
| Cached read tokens | $0.04 per 1M tokens | Whether and when cached input is reported and billed at this rate. |
| Context window | Up to 1M tokens | TokenRa notes that current account limits apply. |
Source: the TokenRa Ox Alpha model page, which also states: “Check the console for current rates and account limits.” Record the date and provider whenever you quote a rate, because a price is a time-bound fact.
Turn per-token rates into a request cost.
Estimate a request as (input tokens ÷ 1,000,000 × $0.14) + (output tokens ÷ 1,000,000 × $0.52). The table applies that formula to a few request sizes.
| Request | Input tokens | Output tokens | Cost |
|---|---|---|---|
| Short coding question | 2,000 | 500 | $0.00054 |
| Multi-file review | 50,000 | 2,000 | $0.00804 |
| Large-context analysis | 500,000 | 4,000 | $0.07208 |
| Near-full 1M-token context | 1,000,000 | 4,000 | $0.14208 |
| 100 requests like the one above | 100,000,000 | 400,000 | $14.208 |
These examples are arithmetic at the rates above, not a quote or a forecast. They price every input token at the standard rate; cached-read pricing, retries, failed requests, and rate-limit behavior change the total. When the route returns token counts, use them to measure your own workload, as described in the API reference, and budget from observed completed requests rather than from one prompt.
A funded TokenRa account is required.
Pricing is only part of the cost of trying a model. TokenRa’s Free AI Models page describes how account funding works:
- New accounts start with a $0 balance, and TokenRa states that it gives no free signup credit.
- The minimum top-up is $10, and TokenRa states that credits do not expire while the account is active.
- In TokenRa’s own test on 31 October 2026, a request from a $0-balance account was rejected with HTTP 403 and the error code
insufficient_user_quota, even for a $0 model.
These are TokenRa’s statements as of that page’s check date, not terms set by this site. Confirm them in the TokenRa console before you plan a budget.
Other routes have their own prices and availability.
The same model label can appear on more than one provider. Treat each route as a separate price list and availability record, and do not carry a price from one route to another.
| Route | What the public listing showed |
|---|---|
| TokenRa | The per-token rates listed above. Confirm current rates in the console. |
| OpenRouter: stealth/ox-alpha | The listing is still online and states that the stealth model was developed and operated by ZAI and revealed to be ZAI GLM-5.3-Flash. OpenRouter’s public endpoints API returned no active endpoints for it. |
| OpenRouter: z-ai/glm-5.3-flash | The listing OpenRouter points to is a metered model whose price is set by each provider endpoint. Check its live listing. |
The OpenRouter route guide documents the earlier setup. Verify the current state of any route before you depend on it.
Keep spend predictable.
- Cap output. Set an output limit that matches the task, for example
max_tokenswhere the route accepts it, instead of leaving it open. - Send only the context the task needs. A 1M-token window is a ceiling, not a default. Every input token is billed at the input rate unless the route applies cached-read pricing.
- Measure real usage. Log token counts, status, and request IDs for every call, and budget from observed completed requests.
- Set your own limits. Enforce an application-level budget, per-user quotas, and alerts instead of relying on the provider balance alone.
- Plan for failure. Handle balance, quota, and rate-limit errors with a visible fallback so a price or availability change does not become an outage.
- Recheck on a schedule. Compare the live TokenRa listing with the rate in your budget model and record the date.
Ox Alpha pricing questions.
How much does the Ox Alpha API cost?
On 1 November 2026, TokenRa listed Ox Alpha at $0.14 per 1M input tokens, $0.52 per 1M output tokens, and $0.04 per 1M cached read tokens. Check the TokenRa console for current rates and account limits.
Is Ox Alpha free?
Not on the TokenRa listing reviewed here. Ox Alpha is billed per token, and TokenRa’s Free AI Models page lists a different model at $0. LiveBench’s $0.000 cost-per-successful-task value is a benchmark field, not a price.
Do I need to add funds before I can call the API?
TokenRa states that new accounts start with a $0 balance, that it gives no free signup credit, and that the minimum top-up is $10. Confirm the current terms in the TokenRa console.
Are the prices the same on OpenRouter?
No. Each provider sets its own prices and availability. On 1 November 2026, OpenRouter’s public endpoints API returned no active endpoints for stealth/ox-alpha, so check the live listing of any route you plan to use.
How do I check the current price?
Open the TokenRa console or the public Ox Alpha model page, record the date, and compare the listed rate with the token counts returned by your own requests.