Qwen OAuth Free Tier Policy Adjustment

[pomelo-nwu]

What would you like to be changed?

Adjust the Qwen OAuth free tier policy with the following changes:

1. Reduce daily free quota: Change from 1,000 requests/day to 100 requests/day effective immediately
2. Phase out free tier: Completely close the Qwen OAuth free entry point on 2026-04-15

Why is this needed?

This is a product strategy adjustment to better manage the free tier usage and costs.

Consider using OpenRouter, Fireworks AI, or Alibaba Cloud ModelStudio as alternatives.

[Maddock-DR]

hi qwencode的Bro，我发现阿里云百炼有免费额度可以使用。https://bailian.console.aliyun.com/cn-beijing?tab=doc#/doc/?type=model&url=2766612

[Ttungx]

终止免费后还会有人用吗🤔

[Wizard254]

TL;DR:

• Kill the free plan (with such reduced limits) and introduce Lite plans (from $10) in the coding plan; loose in the long run due to no cheap training data.
• Keep the free plan at ~500 req/day and keep the coding plan at $50.

---

Just a recommendation.

The free tier plays a crucial role to some fanatic developers, especially those that are in or just out of college/university and still don't have a justifiable income stream, to build POC projects for showcase.

This also creates an imbalance in Qwen's profit margin if too many users are on the free plan than on the subscription plan; but, in the long run will improve the model as free tier interactions can be used to upgrade the model.

Restricting this to 100 requests, from 1000 requests, is a steep change; and will most likely not suffice for even POC projects; this will most likely kill the free tier which was initially at 2000 requests.

With 100 requests; users will most likely just ask qwen to perform a task, then, because they run out of quota, can't ask follow up questions. Quality training data is on the follow -up questions; where the user is given a chance to correct the model and steer it in the right direction; because the assumptions the model made in the first ask were flawed.

The OP recommends the Qwen Coding Plan, which, also checking closely, starts at $50; for starters (university graduates alike), this is a near impossible upgrade (price), and also overkill (requests/month). Thus, a consideration of the re-introduction of the Lite Plans (from $10) would be a great option.

If $50 ==> 90000 requests
Then $10 ===> 18000 requests (~600 requests per day)

Note: I'm not sure if the above math works in production (it's theoretical); and why the initial lite plan was removed, in the first place.

In the near future, LLM models will only be as good as the training data they can get, which is only cheaply available on the free tier.
If every body is on a paid plan (which must be an automatic opt out of training), then, the model risks falling behind; or investing heavily on training data from the likes of https://scale.com.

The above is the bet Google must have made with Antigravity; luring free tier users to use the overpriced Claude Opus models and paying the expensive API costs to Anthropic, just to get quality (real world) training data.

Key:
POC => Proof of Concept
1 month => 28 days

[GSTbyMM]

Please don't do this. Hand-folded request. I am in the middle of a big project.

[raccoonwannafly]

@Wizard254 no point persuading them its the order from higher up cutting budget lol