Qwen API Pricing (November 2026)
Qwen API prices run from $0.03 to $7.50 per million tokens across 18 current models. The cheapest is Qwen3.7 Flash at $0.03 input and $0.13 output per million tokens. The flagship, Qwen3.8 Max, costs $2.00 input and $6.00 output per million tokens. Every rate below was read on Alibaba (Qwen)'s own pricing page, on the date shown in its row.
| Model | Input /M | Cache read /M | Output /M | Batch in /M | Batch out /M | Context | Status | Checked |
|---|---|---|---|---|---|---|---|---|
Qwen3.8-27B qwen3.8-27b | $0.50 | — | $3.00 | — | — | 1M | Current | 2026-10-29 |
Qwen3.8-2.4T-A95B qwen3.8-2.4t-a95b | $2.00 | — | $6.00 | — | — | 1M | Current | 2026-10-29 |
Qwen3.8 Max qwen3.8-max | $2.00 | — | $6.00 | — | — | 1M | Flagship | 2026-10-29 |
Qwen3.8-Flash qwen3.8-flash | $0.15 | — | $0.47 | — | — | 1M | Current | 2026-10-29 |
Qwen3.8-Omni-Flash qwen3.8-omni-flash | $0.15 | $0.016−89% | $0.47 | — | — | 1M | Current | 2026-10-29 |
Qwen3.7 Max qwen3.7-max | $2.50 | — | $7.50 | $1.25 | $3.75 | 1M | Current | 2026-10-29 |
Qwen3.7 Plus qwen3.7-plus | $0.40 | — | $1.60 | $0.20 | $0.80 | 1M | Current | 2026-10-29 |
Qwen3.7 Flash qwen3.7-flash | $0.03 | — | $0.13 | $0.015 | $0.065 | 1M | Current | 2026-10-29 |
Qwen3.6 35B A3B qwen3.6-35b-a3b | $0.375 | — | $2.25 | — | — | 256K | Current | 2026-10-30 |
Qwen3.6 27B qwen3.6-27b | $0.60 | — | $3.60 | — | — | 256K | Current | 2026-10-30 |
Qwen3.6 Plus qwen3.6-plus | $0.50 | $0.05−90% | $3.00 | — | — | 1M | Current | 2026-10-30 |
Qwen3.6 Flash qwen3.6-flash | $0.25 | $0.025−90% | $1.50 | — | — | 1M | Current | 2026-10-30 |
| Qwen 3.5 397B A17B | $0.60 | — | $3.60 | — | — | 256K | Current | 2026-10-29 |
| Qwen 3.5 122B A10B | $0.40 | — | $3.20 | — | — | 256K | Current | 2026-10-29 |
| Qwen 3.5 Plus | $0.40 | — | $2.40 | — | — | 1M | Current | 2026-10-29 |
| Qwen 3.5 Flash | $0.10 | — | $0.40 | — | — | 1M | Current | 2026-10-29 |
| Qwen3 VL Plus | $0.20 | — | $1.60 | — | — | 256K | Current | 2026-10-29 |
Qwen3 Coder Flash qwen3-coder-flash | $0.30 | — | $1.50 | $0.15 | $0.75 | 1M | Current | 2026-10-29 |
| QwQ 32B | $0.287 | — | $0.861 | — | — | 131K | Deprecated | 2026-07-11 |
| Qwen 2.5 Coder 32B Instruct | $0.287 | — | $0.861 | — | — | 131K | Deprecated · → Qwen3.7 Plus | 2026-10-29 |
| Qwen Max (2.5) | $1.60 | — | $6.40 | — | — | 32K | Legacy · → Qwen3.7 Max | 2026-10-29 |
| Qwen Plus (2.5) | $0.40 | — | $1.20 | — | — | 1M | Legacy · → Qwen 3.5 Plus | 2026-10-29 |
| Qwen Flash (2.5) | $0.05 | — | $0.40 | — | — | 1M | Legacy · → Qwen 3.5 Flash | 2026-10-29 |
US dollars per million tokens, list rates on Alibaba (Qwen)'s own API. Cache read is the price of a cached input token. Batch is the asynchronous API, where the vendor offers one. The date in each row is when we last read it on the vendor's pricing page; it links to that page.
Qwen API monthly cost estimates
What light use (10M input + 2M output tokens a month), medium use (100M input + 20M output tokens a month), heavy use (1B input + 200M output tokens a month) costs on each current model, at list price with no caching and no batch discount.
| Model | Input / output per 1M | Light · 10M + 2M | Medium · 100M + 20M | Heavy · 1B + 200M |
|---|---|---|---|---|
| Qwen3.7 Flash | $0.03 / $0.13 | $0.560 | $5.60 | $56.00 |
| Qwen 3.5 Flash | $0.10 / $0.40 | $1.80 | $18.00 | $180 |
| Qwen3.8-Flash, Qwen3.8-Omni-Flash | $0.15 / $0.47 | $2.44 | $24.40 | $244 |
| Qwen3 VL Plus | $0.20 / $1.60 | $5.20 | $52.00 | $520 |
| Qwen3.6 Flash | $0.25 / $1.50 | $5.50 | $55.00 | $550 |
| Qwen3 Coder Flash | $0.30 / $1.50 | $6.00 | $60.00 | $600 |
| Qwen3.6 35B A3B | $0.375 / $2.25 | $8.25 | $82.50 | $825 |
| Qwen3.7 Plus | $0.40 / $1.60 | $7.20 | $72.00 | $720 |
| Qwen 3.5 Plus | $0.40 / $2.40 | $8.80 | $88.00 | $880 |
| Qwen 3.5 122B A10B | $0.40 / $3.20 | $10.40 | $104 | $1,040 |
| Qwen3.8-27B, Qwen3.6 Plus | $0.50 / $3.00 | $11.00 | $110 | $1,100 |
| Qwen3.6 27B, Qwen 3.5 397B A17B | $0.60 / $3.60 | $13.20 | $132 | $1,320 |
| Qwen3.8-2.4T-A95B, Qwen3.8 Max | $2.00 / $6.00 | $32.00 | $320 | $3,200 |
| Qwen3.7 Max | $2.50 / $7.50 | $40.00 | $400 | $4,000 |
Prompt caching and the Batch API cut these figures; the next sections show by how much. For your own token counts, use the LLM API cost calculator.
Qwen API price history
Qwen API price changes and releases
Major releases and the open-weight cadence from Alibaba's Tongyi Lab. Sourced from Qwen blog/changelog, Alibaba Cloud Model Studio, and Wikipedia.
qwen3.8-max-preview ID is replaced by qwen3.8-max at $2/$6 per M, its first published per-token rate, undercutting Qwen3.7 MaxRetired Qwen models and their replacements
25 Qwen models no longer answer on Alibaba (Qwen)'s API. The price is the last one Alibaba (Qwen) listed; the replacement is the current end of the vendor's own succession chain.
| Model | Status | Last price in / out per 1M | Use instead |
|---|---|---|---|
| Qwen3 235B A22B Thinking 2507 | Retired Oct 10, 2026 | $0.23 / $2.30 | Qwen3.7 Plus |
| Qwen3 235B A22B Instruct 2507 | Retired Oct 10, 2026 | $0.23 / $0.92 | Qwen3.7 Plus |
| Qwen3 235B A22B | Retired Oct 10, 2026 | $0.70 / $2.80 | Qwen3.7 Plus |
| Qwen3 Next 80B A3B Thinking | Retired Oct 10, 2026 | $0.15 / $1.20 | Qwen3.7 Plus |
| Qwen3 Next 80B A3B Instruct | Retired Oct 10, 2026 | $0.15 / $1.20 | Qwen3.7 Plus |
| Qwen3 32B | Retired Oct 10, 2026 | $0.16 / $0.64 | Qwen3.7 Plus |
| Qwen3 30B A3B Thinking 2507 | Retired Oct 10, 2026 | $0.20 / $2.40 | Qwen3.7 Plus |
| Qwen3 30B A3B Instruct 2507 | Retired Oct 10, 2026 | $0.20 / $0.80 | Qwen3.7 Plus |
| Qwen3 30B A3B | Retired Oct 10, 2026 | $0.20 / $0.80 | Qwen3.7 Plus |
| Qwen3 14B | Retired Oct 10, 2026 | $0.35 / $1.40 | Qwen3.6 Flash |
| Qwen3 8B | Retired Oct 10, 2026 | $0.18 / $0.70 | Qwen3.6 Flash |
| Qwen3.6 Max Preview | Retired Oct 10, 2026 | $1.30 / $7.80 | Qwen3.7 Max |
| Qwen3 Max | Retired Oct 10, 2026 | $1.20 / $6.00 | Qwen3.7 Max |
| Qwen3 Coder Plus | Retired Oct 10, 2026 | $1.00 / $5.00 | Qwen3.7 Plus |
| Qwen3 Coder Next | Retired Oct 10, 2026 | $0.30 / $1.50 | Qwen3.7 Plus |
| Qwen3 VL Flash | Retired Oct 10, 2026 | $0.05 / $0.40 | Qwen3.6 Flash |
| Qwen 2.5 VL 72B Instruct | Retired May 13, 2026 | $2.80 / $8.40 | Qwen3 VL Plus |
| Qwen 2.5 72B Instruct | Retired May 13, 2026 | $1.40 / $5.60 | Qwen3.7 Plus |
| Qwen 2.5 32B Instruct | Retired May 13, 2026 | $0.70 / $2.80 | Qwen3.7 Plus |
| Qwen 2.5 14B Instruct | Retired May 13, 2026 | $0.35 / $1.40 | Qwen3.6 Flash |
| Qwen 2.5 7B Instruct | Retired May 13, 2026 | $0.175 / $0.70 | Qwen3.6 Flash |
| Qwen VL Max | Retired Oct 10, 2026 | $0.80 / $3.20 | Qwen3 VL Plus |
| QwQ Plus | Retired Oct 10, 2026 | $0.80 / $2.40 | Qwen3.7 Plus |
| Qwen VL Plus | Retired Oct 10, 2026 | $0.21 / $0.63 | Qwen3.6 Flash |
| Qwen Turbo | Retired Oct 10, 2026 | $0.05 / $0.20 | Qwen 3.5 Flash |
Model Studio (Bailian / DashScope) — managed API for the full Qwen catalog, including proprietary tiers. The International (Singapore) endpoint serves non-China traffic; pricing is tiered by input length.
OPEN-WEIGHT · SELF-HOSTMost Qwen tiers ship Apache 2.0 open weights on Hugging Face and ModelScope — fully self-hostable. The most-downloaded open model family, with 300+ models released.
CONSUMER · WEBchat.qwen.ai — free consumer assistant covering chat, vision, and coding. For end-user usage; no API access on this surface.
CHINA · DOMESTICThe China-region Model Studio endpoint keeps data inside mainland China for domestic compliance — the other half of Qwen's dual-deployment model alongside Singapore.
DeepSeek
Low cost-per-token at frontier reasoning quality with MIT open weights, and V4.1 Flash added native vision in September 2026. But a far narrower lineup than Qwen's full-modality catalog, with no audio or video model.
PRIMARY RIVALOpenAI
Frontier benchmarks + ecosystem (GPT-5, ChatGPT, Azure). But closed weights, US data residency, and multiples higher cost than Qwen's open tiers.
FRONTIER LEADERAnthropic
Leads coding + reasoning with Claude and zero-retention defaults. But no open weights and no low-cost / multilingual story to match Qwen.
Meta
Llama open weights with a Western governance posture. But fewer model sizes and modalities than Qwen, and no managed low-cost API.
Mistral
EU-native open weights (Apache 2.0) with GDPR-first residency — the European answer to Qwen's openness, but a smaller catalog and no China-domestic option.
GLM ARCHITECTZhipu (Z.ai / GLM)
The Beijing lab behind the GLM models, also operating internationally as Z.ai. Spun out of Tsinghua University in 2019, Zhipu ships strong coding/agentic models, multiple free tiers, and open weights — and was the first of China's "Six Tigers" to pursue an IPO. It also sits on the US Entity List.
Qwen API pricing FAQ
Practical questions about Qwen pricing, open weights, and deployment.
Q · 01 Which Qwen model should I start with?
+
$0.15/$0.47) as a multimodal daily driver — one rate across the whole 1M window — or Qwen3.8 Max ($2/$6) for the proprietary frontier. Want to self-host? Use the open-weight Qwen3.8-27B or Qwen 3.5 397B. Qwen 3.5 Flash ($0.10/$0.40) is still the cheapest flat-band 1M row if you only need text. See the use case picker. Q · 02 Are Qwen models really open-weight?
+
Q · 03 Why are the prices here lower than what I'm billed?
+
0–32K input range on the International (Singapore) region; longer prompts cost more (for example Qwen3 Coder Plus climbs from $1/$5 to $6/$60 at the 256K–1M tier). Batch API offers a 50% discount on most rows — but not all: the Qwen3.8-Flash row carries no batch badge at all. Some newer models also drop the tiers entirely and bill one rate to 1M. Q · 04 Where is my data stored?
+
Q · 05 How many languages does Qwen support?
+
Q · 06 What's the difference between Qwen3 Max and the open Qwen 3.5 models?
+
$1.2/$6). Alibaba took it offline on October 10, 2026 and points users to Qwen3.7 Max; the top tier today is Qwen3.8 Max. The Qwen 3.5 open MoE models (397B-A17B, 122B-A10B) are Apache-2.0 and self-hostable, at lower per-token cost. Choose a Max tier for the managed top end; choose 3.5 open for control and self-hosting. Q · 07 How does Qwen compare to DeepSeek and Llama?
+
Qwen — branded Tongyi Qianwen (通义千问) — is the large-model family from Alibaba Cloud's Tongyi Lab, which was established in 2022; the Qwen series was first released in 2023 and open-sourcing began that year. It is developed in Hangzhou under Alibaba Group, a public company (NYSE: BABA, HKEX: 9988).
Qwen's defining strategy is open weights at scale. The lab has released 300+ models spanning text, coding, vision, speech, and video, mostly under the permissive Apache 2.0 license on Hugging Face and ModelScope. By March 2026 the family had been downloaded over 940 million times and spawned 200,000+ derivative models — making Qwen the most-downloaded open-weight model family in the world.
Technically, Qwen3 (April 2025) was trained on roughly 36 trillion tokens across 119 languages and dialects, with both dense and mixture-of-experts (MoE) variants from 0.6B up to 397B parameters. The 2026 Qwen 3.5 generation continued the open MoE line (397B-A17B, 122B-A10B) alongside managed Plus/Flash tiers, while the top Qwen3.8 Max flagship is served as a proprietary, cloud-only model.
Distribution is dual-track: managed inference through Alibaba Cloud Model Studio (with separate International/Singapore and China-domestic endpoints), and self-hosting via open weights. Model Studio pricing is tiered by input length — the headline rates here are the 0–32K range on the Singapore region, and costs scale up for longer prompts.
The trade-offs: most models are text-or-vision with strong multilingual coverage, but the company is China-based, so the direct API stores data per its region (China or Singapore), and US export-control / procurement questions apply. Versus DeepSeek, Qwen offers a far broader lineup and multimodal range; versus OpenAI and Anthropic, it trades some frontier-benchmark lead for radically more open weights and lower cost.