Unlimited Basic improves: more tokens, same monthly price

Unlimited Basic improves for the same €14.95/month, including VAT. Qwen 3.8 27B and Qwen 3.6 35B-A3B keep their unlimited usage, subject to fair use, without consuming AI Credits. Your 5,000 monthly credits now also cover up to 13.6 times more tokens with DeepSeek V4.1 Flash.

We have reduced the credit consumption of several premium models and lowered our public API prices. For Unlimited Basic users, this means more room to send context, review code, have conversations and work with agents within the same subscription.

This improvement builds on the new models and Harness access recently added to Unlimited Basic. You keep both Qwen models for regular work and get more premium tokens for tasks where you want to use DeepSeek or GLM.

How many premium tokens your 5,000 AI Credits now cover

The monthly allowance remains 5,000 AI Credits shared across nine premium models. With lower per-token credit consumption for DeepSeek and GLM 5.3, you can process more content before using it up.

Here are examples of dedicating the 5,000 AI Credits to a single premium model, assuming 90% input tokens and 10% output tokens. Each figure shows the combined token total:

Each row is a separate scenario; the amounts are not additive. You can spread your credits across models, but they all draw from the same pool. These are token-volume estimates rather than a guaranteed number of questions or tasks. Actual consumption depends on context, responses and the calls you make.

These figures describe premium model usage with AI Credits. Both Qwen models keep their unlimited usage subject to fair use and do not draw from this pool.

Input tokens include the instructions, documents and history you send with each request. When an agent resends the same context across multiple steps, that content counts again. Output tokens are those counted during the model's generation.

DeepSeek V4.1 Flash: around 13.6 times more tokens from the same allowance

DeepSeek V4.1 Flash drops from 375 to 27.6 AI Credits per million input tokens, and from 1,500 to 110.4 per million output tokens. With the same input-to-output mix, your credits cover approximately 13.6 times the previous token volume.

With the 90% input and 10% output usage mix, each million combined tokens consumes approximately 35.88 AI Credits. The 5,000 monthly credits therefore cover around 139.3 million total tokens with this model.

DeepSeek V4 Flash 0731 also improves: it now consumes 41.4 credits per million input tokens and 82.8 per million output tokens, down from 129 and 258. This allows approximately 3.1 times more tokens from the same credits.

GLM 5.3 drops to 901.6 credits per million input tokens and 2,833.6 per million output tokens. GLM 5.3 Flash keeps its rates of 110 and 440 credits per million, respectively.

You can explore the updated examples for questions, conversations and programming tasks in the Unlimited Basic usage estimates.

Two unlimited Qwen models and more room for premium models

You can use both Qwen models throughout your regular sessions, subject to fair use, and dedicate your credits to DeepSeek, GLM, embeddings, transcription or decisions when you need them.

The price remains €14.95/month, and the credit pool renews with each monthly cycle. If you use up your premium credits, both Qwen models remain available under the plan's conditions. See Unlimited Basic for rate, concurrency and availability conditions.

You can connect the plan to QDivZero Harness, your application or a tool that supports an OpenAI-compatible API. For included usage, use the model identifiers shown in Unlimited inside the platform.

We have also lowered public API prices

The reductions also apply to pay-as-you-go usage with account balance. These are the previous and new rates, expressed in euros per million tokens, with input and output priced separately:

Rates checked on 7 November 2026 against our public endpoint catalog. Amounts and percentages in the table are rounded. See public model pricing for current rates.

For DeepSeek V4.1 Flash, for example, 10 million total tokens drop from around €6.34 to around €0.40 with pay-as-you-go billing, using the same 90% input and 10% output mix. GPT-6 Luna and GPT-6 Sol also halve their rates and are available using account balance.

Unlimited AI Credits and euro account balance are managed separately. The plan's token table uses subscription credit rates; the API pricing table uses the monetary rates of public models. You can combine both options in the same account.

Start with a real task

Open QDivZero, go to Unlimited and review or activate Unlimited Basic. Copy the model identifier from that section, then connect it using your API key and the base URL https://api.qdiv0.com/v1.

Start with the Qwen 3.8 27B API guide or prepare an integration using the inference documentation. Choose a familiar task and see how much context, generation and work your plan can cover now.