Qwen OpenRouter vs Official Pricing (2026)

For Qwen3-Max, the OpenRouter-routed price of $0.78 in / $3.90 out per million tokens is cheaper than Alibaba's own International endpoint, which runs about $2.40 / $12.00 on the 32K-128K input tier, so the wrapper undercuts the source. That inversion is the whole story for international buyers, and it has two sharp exceptions worth knowing before you commit.

Most buyers assume the model maker always sells its own model cheapest, then a reseller adds margin on top. With Qwen3-Max the opposite holds today. The routed price we see in the live OpenRouter catalog sits well below Alibaba's published International rate for the flagship. The practical upshot is that "go direct to the vendor" is not automatically the cheaper path here, and the gap is large enough to change a budget.

Qwen3-Max: routed vs official per tier (verified 2026-06)

Here is the comparison that matters, flagship against flagship. The routed column is the OpenRouter-routed price; the official column is Alibaba's International rate for the same model, which is tiered by how large your input prompt is.

According to Alibaba Cloud Model Studio pricing, the International Qwen3-Max rate is tiered by input-prompt size, with the 32K-128K bracket landing near $2.40 in and $12.00 out per million tokens. We list that as pending native re-verification because in this environment Qwen was reached through OpenRouter, not a native Alibaba key, so the official figures come from sourced research rather than a billed call we placed ourselves. The routed price, by contrast, is one we can confirm against real billed usage.

We measured Qwen3-Max via OpenRouter (first-hand evidence)

Published rate cards are one thing; what the meter actually reads is another. So we called Qwen3-Max through OpenRouter and recorded the billed cost, not an estimate. On a short general-purpose prompt the run consumed 38 input tokens and 63 output tokens and billed $0.00027534, returning in 3.22 seconds. That call was routed via OpenRouter, not placed on a native Alibaba endpoint, and we disclose that because the routing is exactly what makes the price competitive.

Back the measured figure out to the rate it implies and it lines up with the $0.78 / $3.90 routed card, not the roughly $2.40 / $12.00 International tier. A direct Alibaba call on the higher tier would have cost several times more for the same tokens. That is the buyer insight in one data point: for this model, the route you pick moves the bill more than the prompt you send.

When OpenRouter wins, and when official wins

The routed price is not a free lunch in every case. Two factors flip the decision back toward Alibaba's own endpoint, and both are about something other than per-token cost.

The first is free quota. According to Alibaba Cloud Model Studio pricing, the International Singapore deployment ships new accounts a free token allowance of roughly one million tokens for the first 90 days, while the US (Virginia) Global deployment carries no such free quota. If you are prototyping and your volume fits inside that allowance, the effective cost on Singapore is zero for three months, which no routed per-token price can beat. The routed price wins on steady production volume; the free quota wins on early, low-volume experimentation.

The second is data residency and compliance. A routed call passes through a third-party aggregator's infrastructure before it reaches the model, which adds a party to your data path. Buyers under strict residency or vendor-of-record rules often cannot accept that, and for them the question is not which price is lower but which contract their legal team will sign. Alibaba's own Singapore or Virginia endpoints give a direct vendor relationship and a known region, and that is worth paying the higher per-token rate for.

Qwen is a Chinese large language model family from Alibaba that is reachable to US developers through Alibaba Cloud Model Studio International, so US access is not the blocker some buyers assume. The real decision is route economics versus contractual control.

How this compares to DeepSeek

It helps to anchor Qwen against another China-built option. According to DeepSeek API Docs, DeepSeek V4-Flash bills $0.14 in / $0.28 out per million tokens on its official endpoint, far below even the routed Qwen3-Max price. The two are not rivals for the same job: V4-Flash is a budget workhorse, while Qwen3-Max is a flagship-class model you reach for when you want more capability per call.

One asymmetry to keep honest: our DeepSeek anchor was measured on DeepSeek's own official endpoint, while every Qwen figure here was routed via OpenRouter. We do not blur those. The DeepSeek number is a native-endpoint price; the Qwen numbers are routed prices. For a deeper DeepSeek breakdown, see the DeepSeek API pricing hub.

FAQ

Is Qwen3-Max really cheaper on OpenRouter than on Alibaba's own endpoint? On current rates, yes for the flagship. The OpenRouter-routed price is $0.78 in / $3.90 out per million tokens, while Alibaba's International 32K-128K input tier runs about $2.40 / $12.00 according to Alibaba Cloud Model Studio pricing. Treat the official figure as pending native re-verification.

Why is the official Alibaba price marked "needs native re-verify"? Because in our environment Qwen was reached through OpenRouter, not a native Alibaba key. The routed prices are billed and confirmed; the official Alibaba International figures come from sourced research and have not yet been re-checked on a live native DashScope key.

When should I use Alibaba's official endpoint instead of a routed price? Two cases. If you are prototyping inside the Singapore deployment's roughly 1M-token free quota for the first 90 days, official is effectively free. If your data-residency or vendor-of-record rules forbid passing calls through a third-party aggregator, official gives you a direct relationship and a known region.

Can US developers access Qwen at all? Yes. Qwen is accessible to US developers through Alibaba Cloud Model Studio International, via the Singapore endpoint or the US (Virginia) Global deployment, in addition to routed access through aggregators.

---

This is part of the Qwen API pricing hub, where the full routed-versus-official rate card lives.

Author: Kevin Fan, Customer Success Manager at China LLM Directory, specializing in Chinese LLM ecosystem pricing. Last verified: 2026-06-26.

<!-- METADATA { "title": "Qwen: OpenRouter vs Official Alibaba Pricing (2026)", "slug": "qwen-openrouter-vs-official-pricing", "meta_description": "Qwen3-Max routed via OpenRouter ($0.78/$3.90) undercuts Alibaba's own International tier (~$2.40/$12.00). When routed wins vs free quota and residency. 2026-06.", "focus_keyword": "qwen openrouter vs official pricing", "secondary_keywords": ["qwen3-max pricing", "alibaba model studio pricing", "qwen openrouter price", "qwen api cost international"], "tags": ["Qwen", "API Pricing", "OpenRouter", "Alibaba Cloud"], "category": "Pricing", "cluster_id": "qwen-api-pricing", "cluster_role": "micro", "hub_slug": "qwen-api-pricing", "evidence_file": "clients/china-llm-aggregator/articles/qwen-api-pricing-evidence.json", "needs_native_reverify": true, "verified_until": "2026-09-24", "author_name": "Kevin Fan", "author_title": "Customer Success Manager", "author_linkedin": "", "author_expertise": ["Chinese LLM ecosystem", "AI infrastructure pricing", "model benchmarking", "cross-border AI compliance"], "faq_pairs": [ {"q": "Is Qwen3-Max really cheaper on OpenRouter than on Alibaba's own endpoint?", "a": "On current rates, yes for the flagship. The OpenRouter-routed price is $0.78 in / $3.90 out per million tokens, while Alibaba's International 32K-128K input tier runs about $2.40 / $12.00 according to Alibaba Cloud Model Studio pricing. Treat the official figure as pending native re-verification."}, {"q": "Why is the official Alibaba price marked needs native re-verify?", "a": "Because in our environment Qwen was reached through OpenRouter, not a native Alibaba key. The routed prices are billed and confirmed; the official Alibaba International figures come from sourced research and have not yet been re-checked on a live native DashScope key."}, {"q": "When should I use Alibaba's official endpoint instead of a routed price?", "a": "Two cases. If you are prototyping inside the Singapore deployment's roughly 1M-token free quota for the first 90 days, official is effectively free. If your data-residency or vendor-of-record rules forbid passing calls through a third-party aggregator, official gives a direct relationship and a known region."}, {"q": "Can US developers access Qwen at all?", "a": "Yes. Qwen is accessible to US developers through Alibaba Cloud Model Studio International, via the Singapore endpoint or the US (Virginia) Global deployment, in addition to routed access through aggregators."} ], "external_links_used": [ {"url": "https://www.alibabacloud.com/help/en/model-studio/model-pricing", "source_name": "Alibaba Cloud Model Studio pricing", "claim": "Qwen3-Max International tiered by input size ~$2.40/$12.00 at 32K-128K; Singapore free quota ~1M tokens for 90 days; Virginia no free quota (needs native re-verify)"}, {"url": "https://api-docs.deepseek.com/quick_start/pricing/", "source_name": "DeepSeek API Docs", "claim": "DeepSeek V4-Flash official price $0.14 in / $0.28 out, used as the China-built comparison anchor"} ], "internal_links_used": [ {"url": "/blog/qwen-api-pricing/", "anchor_text": "Qwen API pricing hub", "type": "hub"}, {"url": "/blog/deepseek-api-pricing/", "anchor_text": "DeepSeek API pricing hub", "type": "cross-cluster"} ], "first_hand_evidence": { "source": "qwen-api-pricing-evidence.json run label qwen3-max_general", "measured": "qwen3-max via OpenRouter: 38 input tokens, 63 output tokens, billed $0.00027534, latency 3.22s; consistent with routed card $0.78/$3.90, not the ~$2.40/$12.00 Alibaba International tier", "routing": "OpenRouter-routed (not a native Alibaba endpoint)", "captured": "2026-06-26" }, "images_status": "spec-only (not generated; FAL_API_KEY unset)", "images": [ {"position": "featured", "type": "generated", "prompt": "Clean editorial comparison diagram of Qwen3-Max pricing: two paths, one labeled OpenRouter-routed $0.78/$3.90 and one labeled Alibaba International tier ~$2.40/$12.00, with the routed path drawn cheaper. Annotate exceptions: free quota and data residency favoring official. Blue and amber palette, minimal background, 16:9.", "alt": "Comparison diagram of Qwen3-Max pricing showing the OpenRouter-routed rate of $0.78 in and $3.90 out undercutting Alibaba's International tier of about $2.40 in and $12.00 out, with free quota and data residency noted as exceptions favoring the official endpoint"} ] } -->