Unlimited GLM 5.3 Flash and Qwen for €39.95 a month with Unlimited Pro

Introducing Unlimited Pro: unlimited GLM 5.3 Flash, Qwen 3.8 27B and Qwen 3.6 35B-A3B, subject to fair use, plus 15,000 monthly AI Credits for ten premium models for €39.95/month, VAT included. Connect them to your application, development tool or QDivZero Harness without preparing a GPU deployment for the included models.

That opens up a practical way of working. One model to review the code you are writing. Another to test an automation with tools. Credits to turn documents into a searchable collection, transcribe audio or try a different solution with DeepSeek, GLM or Kimi. All through one account and an OpenAI-compatible API.

Pro’s value comes from being able to repeat that work. Sending another version of a problem, adding context and checking a solution are part of building with AI. With three models included without per-token charges, you can organise that regular usage around a known monthly fee.

See everything included in Unlimited Pro for €39.95/month.

Why €39.95 a month can make sense for a developer

An AI subscription becomes interesting when you use it for tasks that come back every week. Reviewing an implementation, understanding a library, preparing tests, transforming documents or running an assistant inside your application are good examples.

Spread across twenty working days, €39.95 is approximately €2 per day. That is a way to assess your budget. The subscription and billing are monthly, with VAT included and no minimum commitment.

For that fee, you get three models with unlimited use under fair use and a shared allowance for another ten. Start with models that are already prepared, reuse OpenAI-compatible clients and spend your development time on the integration. Usage included in Pro does not deduct monetary balance.

The difference from Basic helps explain the offer. For an extra €25 a month, Pro makes GLM 5.3 Flash unlimited and increases monthly AI Credits from 5,000 to 15,000. If you use GLM Flash regularly, its requests stop consuming the allowance you can use for other premium models.

The fit depends on your usage. For a few occasional calls, metered access may cost less. For regular work with GLM Flash and Qwen, Pro brings capability and budget together in one plan.

Unlimited GLM 5.3 Flash and Qwen for everyday work

These three models do not use AI Credits in Unlimited Pro. Unlimited use is subject to fair use, availability and the plan’s request and concurrency limits.

GLM 5.3 Flash for coding and agents with tools

A useful agent needs to connect actions. Interpret a request, decide which information is missing, propose a tool and continue with its result. The official GLM 5.3 Flash announcement highlights coding, planning, tool use and automation.

Z.ai reports 84.3 on Terminal-Bench 2.1, 63.4 on DeepSWE v1.1 and 48.8 on AutomationBench v1.0.6. It outperforms GLM 5.2 on those tests. These are creator evaluations using their own environments, rather than QDivZero API measurements or a guarantee for any individual task.

In Pro, you can bring GLM 5.3 Flash into regular work without consuming premium credits. Try giving it a reproducible bug, asking for a plan and letting your application return search or test results. That combination of model, tools and context forms the basis of a coding agent.

Qwen 3.8 27B for adding AI to your application

Qwen 3.8 27B is another option for code and complex instructions. Its official evaluation reports 61.7 on SWE-bench Pro and 89.2 on GPQA Diamond. Those results provide a starting point for choosing what to test in your own project.

Use it as an assistant that explains a codebase, drafts tests or generates an answer from context retrieved by your backend. The Qwen API guide includes Python, history, streaming and tool examples to help build that workflow.

Qwen 3.6 35B-A3B to expand your options

The second Qwen lets you evaluate another answer to the same problem within your subscription. Its official model card reports 73.4 on SWE-bench Verified and 51.5 on Terminal-Bench 2.0.

Compare implementation proposals, explanations or code transformations against your own tests. Benchmark versions and evaluation environments differ, so the figures should be read within their respective publications.

Four AI projects you can start with Unlimited Pro

A coding assistant connected to your repository

Start with a small, checkable task. Explain a function, find where an API response is built or generate tests for an edge case. Your application selects relevant files and sends that context to GLM Flash or one of the Qwen models.

Then add search, file reading and test execution through your agent’s tools. The model proposes actions and your environment executes them with the permissions you define. QDivZero Harness provides an application for working with models, projects and agents.

A search tool that answers questions about your documents

A knowledge base combines two jobs. An embedding model turns content into vectors for retrieving relevant passages. A text model uses those passages to answer a question.

Allocate AI Credits to Qwen3 Embedding 8B and use an unlimited model to write the answer. The Qwen3 Embedding model card describes text and code retrieval, classification and clustering. It can form part of a RAG workflow for technical documentation, catalogs or support information.

Connect those models to your own vector store or explore Retrieval. Additional resources you use retain their own billing conditions.

A workflow that turns audio into useful information

Use credits with Whisper Large V3 and Whisper Large V3 Turbo to transcribe audio. GLM Flash or Qwen can then work from that text to prepare a summary, extract tasks or draft a response for review.

This provides a starting point for meeting notes, an interview archive or a customer-support workflow. The Whisper Large V3 model card documents speech recognition. Transcription and subsequent processing are separate steps you can integrate into your product.

An automation that chooses how to handle each request

Your application can use an unlimited model for regular work and reserve credits for tasks where you want to try DeepSeek, GLM 5.3 or Kimi. Compare an answer, process a document or explore another solution through the same compatible API format.

As your project needs them, LLM Model Routing lets you configure routes and fallbacks, while LLM Firewall adds controls to compatible requests. Model usage follows the access option selected for each destination.

Which models use Unlimited Pro’s 15,000 AI Credits?

Alongside the three unlimited models, Pro includes these ten models sharing one monthly allowance. Here is one way to organise them around the work you can test.

See the official model cards for DeepSeek V4.1 Flash, GLM 5.3, Kimi K3 and Kimi K3.1 Pro for model details. Effective formats, tools and limits depend on the endpoint you use.

Kimi K3.1 Pro and GLM 5.3 consume credits. GLM 5.3 Flash is included without credit usage in Unlimited Pro. That distinction helps you allocate work deliberately.

An example of splitting 15,000 AI Credits across several tasks

Imagine you are preparing a document-search application and want to test several text models during the month. This illustrative budget combines embeddings and queries instead of allocating the full allowance to one model.

The calculation uses plan AI Credit rates checked on November 12, 2026. These are accumulated volumes, rather than a single request. Actual charges depend on how requests are distributed and their billing rounding. Resent context counts again as input.

In this scenario, you would use approximately 37% of the monthly allowance. Both Qwen models and GLM 5.3 Flash would continue without consuming AI Credits, subject to fair use. Audio and decisions have their own usage units.

Credits renew with each monthly cycle and unused credits expire. If you run out, the three unlimited models remain available under the plan’s conditions. Euro balance is managed separately. Check the Pro credit estimates and conditions.

Unlimited Basic or Pro, which should you choose?

Both plans let you connect models to applications without deploying GPUs for included usage. The choice depends on how often you use GLM Flash and how much room you need for models that consume credits.

Basic can fit when your regular work centres on Qwen and you use other models occasionally. Pro becomes more interesting when you want daily GLM Flash access, more model experimentation or a combination of text, embeddings and audio with an allowance three times larger.

Compare the plans and choose Unlimited Pro.

Unlimited Pro is available in your QDivZero account

Activate Unlimited Pro in the subscription section of the QDivZero platform and connect the plan’s models to your applications or QDivZero Harness through the OpenAI-compatible API.

Included models are ready to use without deploying GPUs. Each model has an identifier tied to its billing option: the subscription identifier provides included Pro usage, while a public variant may deduct monetary balance.

The Unlimited documentation covers included access, and the integration documentation explains how to connect your application.

Unlimited Pro, GLM Flash and Qwen FAQ

Is GLM 5.3 Flash unlimited in Unlimited Pro?

Yes. GLM 5.3 Flash is included without per-token billing or AI Credit usage, subject to fair use, plan limits and availability. In Basic, it uses credits from the monthly allowance.

Which Qwen models are unlimited?

Qwen 3.8 27B and Qwen 3.6 35B-A3B. Both are included in Basic and Pro without AI Credit usage, subject to fair use. Qwen3 Embedding 8B uses credits to process input text.

Can I use Pro for coding agents?

Yes. Connect the models to a compatible application or tool. Your agent provides context, tools and action execution. Check which features each model supports and review results against your project’s tests.

Does Pro include a dedicated GPU deployment?

Included models are prepared for API access. Dedicated deployments are a separate option and use balance under their own conditions. You can combine both in the same account.

What happens when the 15,000 AI Credits run out?

Credit-consuming models become available again when the monthly allowance renews. Both Qwen models and GLM 5.3 Flash remain included under fair use. Monetary balance does not automatically replace credits.

Where do Pro models run?

Unlimited models run exclusively in Europe. Models that use AI Credits run in Europe and the United States. Our providers strictly follow a zero data retention (ZDR) policy. Review the security and privacy information for your integration.

Is there a minimum term or a per-user fee?

Pro costs €39.95/month with VAT included, no minimum commitment and no extra per-user fees. Cancel to avoid the next charge. Cancellation does not refund the period already paid.

Choose a task in your project and start there

A bug you want to understand. A document collection that needs to answer questions. A workflow that turns a transcript into prepared work. A task you can test and review is the best starting point for assessing Pro.

For €39.95/month, you get unlimited GLM 5.3 Flash and two Qwen models under fair use, plus 15,000 AI Credits to complete your application with another ten models. One fee for sustained AI work and room to discover which combination is useful to you.

See all Unlimited Pro models · Create your account and choose Pro · Prepare your first integration.