GLM 5.3 Flash API Pricing: Cost Per Million Input and Output Tokens

The short answer

The supplied pricing report lists GLM 5.3 Flash at $0.15 per million input tokens and $0.50 per million output tokens. It also reports cached input tokens at $0.03 per million. At those rates, one million uncached input tokens plus one million output tokens costs $0.65.

That price should not be confused with the separate GLM 5.3 model. OneMux lists GLM 5.3 at $1.05 per million input tokens and $3.30 per million output tokens. Its model catalogue separately lists GLM 5.3 Flash at $0 for input and $0 for output. These are distinct catalogue entries, so the exact model ID and route matter more than the model family name alone.

GLM 5.3 Flash token pricing

The supplied source describes three relevant token categories:

The external source for these figures is MindStudio's GLM 5.3 Flash pricing article. The source summary supports the rates above; it does not establish that every gateway or API route applies the same prices.

Cost examples

For a simple estimate, use this formula:

cost = (input_tokens / 1,000,000 × input_rate)
     + (output_tokens / 1,000,000 × output_rate)

Using the reported GLM 5.3 Flash rates:

• 100,000 input tokens and 20,000 output tokens: $0.015 + $0.010 = $0.025.
• 1 million input tokens and 1 million output tokens: $0.15 + $0.50 = $0.65.
• 10 million input tokens and 2 million output tokens: $1.50 + $1.00 = $2.50.

These examples use uncached input throughout. If part of the prompt qualifies as cached input, substitute the cached rate for that portion rather than charging the standard input rate.

What cached input changes

Cached input matters most when applications repeatedly send a large, stable prefix. Typical examples include a long system instruction, a fixed policy, a product catalogue, or a recurring reference document. The reported cached rate of $0.03 per million tokens is one-fifth of the reported standard input rate of $0.15.

The saving applies only to the cached portion. For example, if a request contains 800,000 cached input tokens and 200,000 uncached input tokens, the input calculation at the reported rates is:

(0.8 × $0.03) + (0.2 × $0.15) = $0.054

Output remains separate. If that request generates 100,000 output tokens, add $0.05, producing a combined estimated cost of $0.104.

Do not assume that repeating identical text automatically makes it cached. Your integration needs to follow the relevant provider or gateway behavior, and your usage records should distinguish cached and uncached tokens when that information is available.

Provider-reported pricing versus OneMux catalogue pricing

The most important purchasing question is not simply “What does GLM 5.3 cost?” It is which GLM 5.3 model and which route will process the request?

OneMux provides access to models through one OpenAI-compatible API and lists separate entries for GLM 5.3 and GLM 5.3 Flash:

The first row is the standard GLM 5.3 entry, not GLM 5.3 Flash. The second row reflects the OneMux catalogue listing supplied for GLM 5.3 Flash. Because the listed OneMux figures differ from the provider-reported figures in the external source, treat them as route-specific catalogue information rather than interchangeable prices.

Before building a budget, check the selected model in the OneMux model catalogue, confirm the request's model identifier, and review the current account and usage details. OneMux is designed to centralize model access, routing, API keys, spend visibility, and credit top-ups through one interface.

Which option fits your workload?

Choose the reported Flash rates when token economics are the priority

The reported $0.15 input and $0.50 output rates make GLM 5.3 Flash attractive for workloads with substantial request volume, such as support classification, lightweight content transformations, structured extraction, and high-frequency application interactions. The correct choice still depends on the behavior and quality your application requires; price alone does not establish suitability.

Consider standard GLM 5.3 when you need that exact model entry

If your integration specifically requires the standard GLM 5.3 catalogue entry, OneMux lists it at $1.05 input and $3.30 output per million tokens. That is materially different from the reported Flash figures, so forecasts should use the rates for the exact deployed model rather than extrapolating from the Flash name.

Use a gateway when operational simplicity matters

For teams managing several providers, OneMux offers a single OpenAI-compatible API for model access and routing. That can reduce the number of provider-specific integrations your developers maintain while giving operators one place to manage keys, monitor spend, and top up credits. See the OneMux quickstart and API documentation for integration details.

Budgeting checklist for developers and operators

1. Record the exact model ID. “GLM 5.3” and “GLM 5.3 Flash” are separate entries in the supplied OneMux catalogue.
2. Separate input from output. Output is reported at a higher per-token rate than input, so long generated responses can dominate spend.
3. Measure prompt repetition. Stable prefixes may create an opportunity for cached-input pricing, but do not budget that saving without confirming eligibility.
4. Estimate by workload shape. A support bot, batch extractor, and long-context research tool can have very different input-to-output ratios.
5. Verify the active route. Provider pricing and gateway catalogue pricing may not match. Review the route and current account pricing before launch.
6. Track spend in production. Set internal thresholds and inspect token usage rather than relying only on request counts.

For broader cost planning, compare the model-specific rates with the information on OneMux pricing. A per-million-token figure is useful only when paired with realistic token volumes, cache behavior, and the actual route used by the application.

Conclusion

The reported GLM 5.3 Flash rates are $0.15 per million input tokens, $0.50 per million output tokens, and $0.03 per million cached input tokens. That makes a one-million-input, one-million-output workload $0.65 before any route-specific considerations.

OneMux presents a separate picture: GLM 5.3 is listed at $1.05 input and $3.30 output, while GLM 5.3 Flash is listed at $0 for both fields in the supplied catalogue. The practical decision is therefore to identify the exact model and route first, then calculate cost from the applicable input, output, and cached-token categories.