NewHow we catch a seller listing the wrong model, and refund you if one does.
UberLLM

Cheapest Qwen API: below Alibaba Cloud on 3 of 3 models (November 2026)

Updated November 20, 2026, 09:00 UTC. UberLLM prices are read live when the page is built (every hour); other providers' prices were checked November 19, 2026 or read from their public price lists at build. UberLLM wrote this page and is one of the providers in it.

On November 20, 2026, UberLLM was below Alibaba Cloud’s own price on both input and output for 3 of the 3 Qwen models we compared, with UberLLM’s card fee added. Among the real-time prices in the table, UberLLM was lowest on 15 of the 15 models. UberLLM prices are the best seller’s price now; they move during the day.

Some UberLLM prices here are more than 70% below list. They are what buyers were billed, not errors. Why so cheap?

ProviderHow you payPrice against Alibaba Cloud’s
UberLLMPer token, prepaid; 5.5% + $0.50 card fee when adding creditBelow it on 3 of 3 models (card fee added)
Alibaba Cloud Model StudioPer token, direct from Alibaba CloudThe list price
OpenRouterPrepaid credit; 5.5% when buying creditAgainst Alibaba Cloud’s price: the same on 2, lower on 1 of 3
Vercel AI GatewayCredit in your Vercel accountAgainst Alibaba Cloud’s price: the same on 3 of 3
RequestyPrepaid; 5% on model costs, added when you top upAgainst Alibaba Cloud’s price: lower on 1, higher on 1 of 2
DeepInfraRuns models on its own GPUs, per tokenAgainst Alibaba Cloud’s price: lower on 2 of 2
Novita AIRuns open models on its own GPUs, per tokenAgainst Alibaba Cloud’s price: the same on 1, lower on 1 of 2

UberLLM first, then Alibaba Cloud’s own API, then the other providers that list these models. The table below shows, model by model, who is lowest. An empty cell means we could not confirm it on the provider's own pages.

Qwen prices, model by model ($ per 1M input / output tokens)

ModelUberLLMAlibaba CloudOpenRouterVercel AI GatewayRequestyDeepInfraNovita AI
Qwen 3.8 27B$0.0002 / $0.0025$0.06 / $2.20$0.06 / $2.20$0.20 / $2.50$0.20 / $2.50$0.42 / $3.00
Qwen3.8 FlashWhy so cheap?$0.015 / $0.047$0.15 / $0.47$0.15 / $0.47$0.15 / $0.47$0.16 / $0.47$0.113 / $0.382$0.15 / $0.47
Qwen 4 MaxWhy so cheap?$0.1522 / $0.4565$1.50 / $4.50$1.425 / $4.275$1.50 / $4.50$1.413 / $4.24$1.413 / $4.24$1.413 / $4.24
Qwen3.8 Omni Flash$0.000168 / $0.000526$0.15 / $0.47$0.15 / $0.47
Qwen3.7 FlashWhy so cheap?$0.0012 / $0.00519$0.03 / $0.13$0.03 / $0.13$0.03 / $0.13
Qwen3.8 Max (0902)$0.0852 / $0.2556$2.00 / $6.00$2.00 / $6.00
Qwen3.8 2.4T A95B$0.0126 / $0.0379$2.00 / $6.00$2.00 / $6.00$2.50 / $6.00$2.00 / $6.00$2.00 / $6.00
Qwen 3.8 Max$0.1136 / $0.3409$2.00 / $6.00$2.00 / $6.00$1.65 / $4.951$2.00 / $6.00
Qwen 3.5 35B A3B$0.0833 / $0.5554$0.14 / $1.00$0.14 / $1.00$0.14 / $1.00$0.25 / $2.00
Qwen 3.5 9B$0.05 / $0.075$0.17 / $0.25$0.09 / $0.13$0.10 / $0.15
Qwen3 Next 80B A3B$0.00086 / $0.00946$0.10 / $1.10$0.09 / $1.10$0.09 / $1.10
Qwen3 30B A3B$0.0838 / $0.3491$0.12 / $0.50$0.12 / $0.50
Qwen3.5 Plus$0.052 / $0.312$0.40 / $2.40
Qwen 3 235B A22B Instruct 2507$0.000567 / $0.00347$0.14 / $0.80
Qwen 3 Coder 480b$0.00297 / $0.0099$0.30 / $1.00

UberLLM: the best seller’s price, read November 20, 2026; the 5.5% card fee is not in the cell. Alibaba Cloud: its own pricing page, checked November 19, 2026; an empty cell means its page does not list that model. OpenRouter, Vercel AI Gateway, Requesty, DeepInfra and Novita AI: each one’s public model list, read when this page was built.

Each provider, good and bad

UberLLM

Pick it if: You want Qwen below Alibaba Cloud’s price, billed per request with no plan, and the same key for every other model.

Watch for: Requests go to third-party providers, not to Alibaba Cloud directly: if you need Alibaba Cloud’s own data terms, use its API. The price is a market price and can be higher when the lower-priced sellers are busy (cap it with provider.max_price).

Source: UberLLM Qwen prices (live, November 20, 2026)

Alibaba Cloud Model Studio

Pick it if: You want Alibaba Cloud’s own service and its 1 million free tokens per model for new Singapore accounts.

Watch for: Some models are priced by prompt length; check the tier your prompts fall in.

Source: Alibaba Cloud pricing (checked 2026-11-19)

OpenRouter

Pick it if: You already use OpenRouter for other models and want automatic failover.

Watch for: Its 5.5% credit fee comes on top of the prices in the table.

Source: OpenRouter pricing, its model list (checked 2026-11-19)

Vercel AI Gateway

Pick it if: Your app runs on Vercel or uses the AI SDK.

Watch for: No markup on provider token prices, so no saving per token either.

Source: Vercel AI Gateway pricing, its model list (checked 2026-11-19)

Requesty

Pick it if: You want routing rules, fallbacks and per-team spend limits.

Watch for: The 5% top-up fee comes on top of the prices in the table.

Source: Requesty pricing, its model list (checked 2026-11-19)

DeepInfra

Pick it if: Steady prices on US hosting for open models.

Watch for: Not every version is listed.

Source: DeepInfra pricing, its model list (checked 2026-11-19)

Novita AI

Pick it if: Low prices on open models.

Watch for: Check that your version is listed.

Source: Novita pricing, its model list (checked 2026-11-19)

How to switch to UberLLM

UberLLM takes OpenAI-style requests: change the base URL, the key and the model id (Qwen ids on UberLLM start with qwen/).

python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.uberllm.dev/v1",  # was: https://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1
    api_key="ull_your_key_here",
)

resp = client.chat.completions.create(
    model="qwen/qwen-3-8-27b",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)

No UberLLM fee on requests. Adding credit by card costs 5.5% + $0.50.

Questions

Is there a free Qwen API?

Alibaba Cloud Model Studio gives 1 million free tokens on models such as Qwen 4 Max and Qwen3.8 Flash, in the Singapore region only, for 90 days.

Does the Qwen API cost more for long prompts?

On some models, yes: Alibaba Cloud bills every token of a request at the price of its prompt-length tier. Qwen3.7 Flash, for example, rises above 32K tokens.

Can I cap what a Qwen request costs on UberLLM?

Yes. Send provider.max_price with the most you will pay per million tokens; the request only goes to sellers at or under it.

Are there other fees on UberLLM?

No UberLLM fee on requests. Adding credit by card costs 5.5% + $0.50.