Qwen3 Max Pricing (2026)

Qwen3-Max routes for about $0.78 per million input tokens and $3.90 output via OpenRouter, while Alibaba's own International endpoint charges roughly $2.40/$12.00 for the 32K-128K bracket, meaning the third-party route is cheaper than the first-party one. That inversion is the part most buyers miss, so it anchors everything below. Every Qwen number we quote here is the OpenRouter-routed value and disclosed as such; the official Alibaba figures are sourced and pending native re-verification.

Qwen3-Max API pricing (verified 2026-06)

Qwen3-Max is Alibaba's flagship Qwen model, the largest general-purpose tier in the family, positioned above Qwen-Plus and the open-weight Qwen3 variants. When buyers ask "what does Qwen3-Max cost," the honest answer depends entirely on which door you walk through, because the two doors are priced very differently.

The OpenRouter-routed price we verified on the live catalog is $0.78 per million input tokens and $3.90 per million output. That is the rate you actually pay when you call the model through OpenRouter, and it is the figure our first-hand testing below was billed against.

According to Alibaba Cloud Model Studio pricing, the official International endpoint prices Qwen3-Max on a tiered scale by input size, landing near $2.40 per million input and $12.00 per million output for the 32K-128K input bracket. That official figure is materially higher than the routed price, and we mark it needs_native_reverify because we have not yet re-checked it on a native DashScope key.

The practical upshot: for the flagship tier, an international buyer routing through OpenRouter pays roughly a third of what Alibaba's own International endpoint charges on the comparable input bracket. That is unusual. For most models the first-party endpoint is the price floor, and third parties add a margin on top. Here the relationship is reversed for the bracket we examined.

Why OpenRouter undercuts Alibaba's own price

The inversion comes down to how each side meters the model. Alibaba's International pricing is tiered by input size, so a long-context request climbs into the $2.40/$12.00 bracket once it crosses 32K tokens. The routed price we measured is flat at $0.78/$3.90 regardless of how the input grows. For any workload that pushes long prompts, the flat routed rate compounds into a large gap rather than a small one.

According to Alibaba Cloud Model Studio pricing, US and international developers reach Qwen through Alibaba Cloud Model Studio International, with a Singapore endpoint that includes a free quota of about one million tokens for 90 days and a US (Virginia) global deployment that carries no free quota. So Qwen is genuinely available to US developers on a first-party key; the question is whether the first-party price is worth paying when the routed price sits lower.

One caution before you treat the routed price as a free lunch. A third-party route adds a dependency on that router's uptime and its own billing layer, and the model served can be a dated snapshot. When we called the coder tier through the router it resolved to qwen3-coder-480b-a35b, a specific pinned build, not an evergreen alias. Budget against the build you actually get.

We measured Qwen3-Max on a live call (first-hand evidence)

Published rate cards tell you the per-million price. They do not tell you what a real call costs or how fast it returns, so we called Qwen3-Max twice through OpenRouter and recorded the billed cost and latency directly from the API response.

On a general-purpose prompt, Qwen3-Max consumed 38 input tokens and produced 63 output tokens, billed at $0.00027534, and returned in 3.22 seconds. On a coding prompt the same model used 36 input and 53 output tokens, billed at $0.00023478, in 2.28 seconds. Both calls were routed via OpenRouter and billed at the $0.78/$3.90 routed rate, not Alibaba's official tiered rate.

What stands out is how small the absolute numbers are at these token counts: well under a third of a cent per short exchange. The coding call came back faster and cheaper than the general one, which tracks with its slightly shorter output. These are tiny single-call figures, useful as a sanity check on the rate card rather than a production projection; multiply by your real token volume to budget.

Qwen3-Max versus DeepSeek on price

If raw cost per token is your only axis, Qwen3-Max is not the cheapest Chinese flagship you can call. According to DeepSeek API Docs, DeepSeek V4-Flash bills $0.14 per million input and $0.28 per million output on its official endpoint. That is roughly 5.5 times cheaper on input and 14 times cheaper on output than the routed Qwen3-Max price, and far cheaper still than Alibaba's official Qwen3-Max bracket.

There is an evidence asymmetry worth stating plainly. We called DeepSeek on its own official endpoint and Qwen via OpenRouter, so the two are not measured through an identical pipe. The DeepSeek anchor cost is computed from its official $0.14/$0.28 rate rather than a live router bill. We disclose this so you weigh the comparison correctly rather than treating it as a head-to-head on one platform. For a deeper DeepSeek breakdown, see the DeepSeek API pricing hub.

The verdict is straightforward. Choose Qwen3-Max when you want Alibaba's strongest general model and the routed price beats the first-party one for your context length. Reach for DeepSeek V4-Flash instead when per-token cost is the dominant constraint and its capability clears your bar, because the gap is large enough to matter at volume.

FAQ

How much does the Qwen3-Max API cost? Via OpenRouter, the verified routed price is $0.78 per million input tokens and $3.90 per million output. Alibaba's official International endpoint charges more, roughly $2.40/$12.00 for the 32K-128K input bracket. We mark the official figure pending native re-verification.

Is Qwen3-Max cheaper through OpenRouter than through Alibaba directly? For the 32K-128K bracket, yes. The OpenRouter-routed $0.78/$3.90 is materially below Alibaba International's roughly $2.40/$12.00 on that bracket. This is unusual; for most models the first-party endpoint is the price floor.

What did a real Qwen3-Max call cost in testing? On our live OpenRouter-routed call, a general prompt of 38 input and 63 output tokens billed $0.00027534 and returned in 3.22 seconds. A coding prompt of 36 in and 53 out billed $0.00023478 in 2.28 seconds.

Can US developers access Qwen3-Max? Yes. According to Alibaba Cloud Model Studio pricing, Qwen is reachable through Model Studio International, including a Singapore endpoint with a 90-day free quota and a US (Virginia) global deployment. Third-party routers also serve it.

Is Qwen3-Max cheaper than DeepSeek? No. DeepSeek V4-Flash bills $0.14/$0.28 officially, far below the routed Qwen3-Max price of $0.78/$3.90. Qwen3-Max competes on capability, not on being the cheapest Chinese flagship per token.

---

This is part of the Qwen API pricing hub, which compares every Qwen tier side by side.

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

<!-- METADATA { "title": "Qwen3-Max API Pricing Explained for 2026 Buyers", "slug": "qwen3-max-pricing", "meta_description": "Qwen3-Max routes for $0.78/$3.90 per 1M via OpenRouter, cheaper than Alibaba's official ~$2.40/$12.00 International bracket. Live test billed $0.00027534.", "focus_keyword": "qwen3-max api pricing", "secondary_keywords": ["qwen3-max cost", "qwen3-max openrouter price", "qwen3 max alibaba pricing", "qwen3-max vs deepseek price"], "tags": ["Qwen", "API Pricing", "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": "How much does the Qwen3-Max API cost?", "a": "Via OpenRouter, the verified routed price is $0.78 per million input tokens and $3.90 per million output. Alibaba's official International endpoint charges more, roughly $2.40/$12.00 for the 32K-128K input bracket. We mark the official figure pending native re-verification."}, {"q": "Is Qwen3-Max cheaper through OpenRouter than through Alibaba directly?", "a": "For the 32K-128K bracket, yes. The OpenRouter-routed $0.78/$3.90 is materially below Alibaba International's roughly $2.40/$12.00 on that bracket. This is unusual; for most models the first-party endpoint is the price floor."}, {"q": "What did a real Qwen3-Max call cost in testing?", "a": "On our live OpenRouter-routed call, a general prompt of 38 input and 63 output tokens billed $0.00027534 and returned in 3.22 seconds. A coding prompt of 36 in and 53 out billed $0.00023478 in 2.28 seconds."}, {"q": "Can US developers access Qwen3-Max?", "a": "Yes. According to Alibaba Cloud Model Studio pricing, Qwen is reachable through Model Studio International, including a Singapore endpoint with a 90-day free quota and a US (Virginia) global deployment. Third-party routers also serve it."}, {"q": "Is Qwen3-Max cheaper than DeepSeek?", "a": "No. DeepSeek V4-Flash bills $0.14/$0.28 officially, far below the routed Qwen3-Max price of $0.78/$3.90. Qwen3-Max competes on capability, not on being the cheapest Chinese flagship per token."} ], "external_links_used": [ {"url": "https://www.alibabacloud.com/help/en/model-studio/model-pricing", "source_name": "Alibaba Cloud Model Studio pricing", "claim": "Qwen3-Max official International tiered pricing ~$2.40/$12.00 for 32K-128K bracket; Model Studio International access via Singapore (90-day free quota) and US Virginia deployment"}, {"url": "https://api-docs.deepseek.com/quick_start/pricing/", "source_name": "DeepSeek API Docs", "claim": "DeepSeek V4-Flash official pricing $0.14 input / $0.28 output per 1M tokens, used as the Qwen-vs-DeepSeek 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 runs qwen3-max_general + qwen3-max_coding", "measured": "qwen3-max general 38in/63out billed $0.00027534 @3.22s; qwen3-max coding 36in/53out billed $0.00023478 @2.28s; all via OpenRouter at $0.78/$3.90 routed rate", "captured": "2026-06-26" }, "images_status": "spec-only (not generated; FAL_API_KEY unset)", "images": [ {"position": "featured", "type": "generated", "prompt": "Clean editorial bar comparison of Qwen3-Max API pricing showing OpenRouter-routed $0.78/$3.90 below Alibaba International official $2.40/$12.00 for the 32K-128K bracket, with DeepSeek V4-Flash $0.14/$0.28 as a lower anchor. Blue and amber palette, minimal background, 16:9.", "alt": "Bar chart comparing Qwen3-Max OpenRouter-routed price of $0.78/$3.90 per million tokens against Alibaba International official $2.40/$12.00 and DeepSeek V4-Flash $0.14/$0.28"} ] } -->