AIHubMixAIHubMix

Qwen Models

140 modelsGeneral models from $0.02/M inputUp to 2M context

Usage

Last 30 days · 2026-07-13 to 2026-08-11

Tokens

202B

Requests

4.3M

Models in use

130 of 140

Tokens per day

038.6B77.3B07-1307-2007-2708-0308-102026-07-13 — 1,455,042,910 tokens2026-07-14 — 1,690,520,375 tokens2026-07-15 — 2,304,259,330 tokens2026-07-16 — 2,122,291,940 tokens2026-07-17 — 3,013,738,415 tokens2026-07-18 — 1,261,102,385 tokens2026-07-19 — 1,236,199,720 tokens2026-07-20 — 3,977,894,690 tokens2026-07-21 — 3,502,157,550 tokens2026-07-22 — 3,594,802,455 tokens2026-07-23 — 5,472,829,720 tokens2026-07-24 — 5,615,339,185 tokens2026-07-25 — 3,836,678,275 tokens2026-07-26 — 3,090,499,560 tokens2026-07-27 — 4,519,148,930 tokens2026-07-28 — 4,549,309,250 tokens2026-07-29 — 3,810,921,310 tokens2026-07-30 — 4,044,367,870 tokens2026-07-31 — 6,050,270,600 tokens2026-08-01 — 3,752,925,355 tokens2026-08-02 — 5,181,574,000 tokens2026-08-03 — 5,167,803,105 tokens2026-08-04 — 6,877,120,050 tokens2026-08-05 — 4,927,219,620 tokens2026-08-06 — 6,027,696,305 tokens2026-08-07 — 77,276,482,745 tokens2026-08-08 — 6,429,860,125 tokens2026-08-09 — 5,918,303,370 tokens2026-08-10 — 11,750,787,285 tokens2026-08-11 — 3,694,289,610 tokens

Which models that traffic went to

  1. qwen3.8-max44.3%89.6B
  2. qwen3.8-max-preview16.4%33.1B
  3. qwen3.5-flash9.0%18.2B
  4. qwen3.6-flash4.9%9.8B
  5. qwen3.7-plus3.9%8B
  6. qwen3.6-plus3.1%6.3B
  7. qwen-turbo2.9%5.9B
  8. qwen3.5-35b-a3b2.1%4.2B
  9. 89 more models13.4%27.1B

Share of 202B tokens. 33 models with traffic report no token counts and cannot be ranked here, including qwen2.5-7b-instruct and Qwen/Qwen2.5-VL-72B-Instruct — they are in the request view.

The two views disagree on purpose: a model can take a large share of the calls and a small share of the tokens — many short requests — or the reverse. Which one matters depends on whether your cost is driven by call volume or by prompt length. Measured on AIHubMix over the last 30 days, counting the 140 model IDs listed on this page; traffic routed through upstream-specific IDs that are not in the public catalog is not included.

All 140 Qwen Models

Open in model list
Qwen models on AIHubMix with input and output modalities, context length, maximum output, price per million tokens including cache read and cache write rates, and measured throughput and latency.
Modalities
qwen3-coder-nextTakes text, returns text.2M64K$0.14$0.55/M——16 tok/s0.40 s
qwen3-coder-30b-a3b-instructTakes text, returns text.2M262K$0.20$0.80/M——115 tok/s0.95 s
qwen3-coder-plusTakes text, returns text.1.05M66K$0.54$2.16/M$0.11/M—83 tok/s1.63 s
qwen3.6-plus-preview-freeTakes text, returns text.1M66KFreeFree/M————
qwen3.7-plusTakes text, returns text.1M64K$0.28$1.13/M$0.06/M$0.35/M41 tok/s2.75 s
qwen3.7-maxTakes text, returns text.1M64K$1.69$5.07/M$0.17/M$2.11/M50 tok/s1.07 s
qwen3.5-flashTakes text, vision, video, returns text.991K64K$0.03$0.28/M$0.0028/M$0.04/M95 tok/s1.25 s
qwen3.5-35b-a3bTakes text, vision, video, returns text.991K64K$0.06$0.45/M——71 tok/s1.27 s
qwen3.5-27bTakes text, vision, video, returns text.991K64K$0.08$0.68/M——71 tok/s1.27 s
qwen3.5-plusTakes text, vision, video, returns text.991K64K$0.11$0.66/M$0.01/M$0.14/M31 tok/s1.92 s
qwen3.5-122b-a10bTakes text, vision, video, returns text.991K64K$0.11$0.90/M——94 tok/s14.59 s
qwen3.5-397b-a17bTakes text, vision, video, returns text.991K64K$0.16$0.99/M——21 tok/s1.75 s
qwen3.6-flashTakes text, vision, video, returns text.991K64K$0.17$1.01/M$0.02/M$0.21/M101 tok/s1.61 s
qwen3.6-plusTakes text, vision, video, returns text.991K64K$0.28$1.69/M$0.03/M$0.35/M34 tok/s2.63 s
qwen3.7-flashTakes text, returns text.991K64K$0.28$1.13/M$0.06/M$0.35/M81 tok/s1.36 s
qwen3.8-max-previewTakes text, vision, returns text.984K131K$0.34$1.01/M$0.07/M$0.42/M48 tok/s2.40 s
qwen3.8-maxTakes text, returns text.984K128K$1.69$5.07/M$0.17/M$2.11/M37 tok/s3.04 s
qwen3.6-35b-a3bTakes text, vision, video, returns text.262K64K$0.25$1.52/M——205 tok/s1.11 s
qwen3-235b-a22b-instruct-2507Takes text, vision, returns text.262K—$0.28$1.12/M——96 tok/s1.03 s
qwen3-235b-a22b-thinking-2507Takes text, vision, returns text.262K—$0.28$2.80/M——87 tok/s0.27 s
qwen3.6-27bTakes text, vision, video, returns text.262K64K$0.42$2.53/M——25 tok/s0.74 s
qwen3-maxTakes text, vision, returns text.262K66K$0.45$1.80/M$0.09/M$0.56/M45 tok/s1.30 s
qwen3-coder-480b-a35b-instructTakes text, returns text.262K—$0.82$3.28/M——1655 tok/s0.92 s
qwen3-coder-flashTakes text, returns text.256K66K$0.14$0.54/M——110 tok/s1.85 s
qwen3-vl-plusTakes text, vision, video, returns text.256K32K$0.14$1.37/M$0.03/M—25 tok/s1.60 s
qwen3-next-80b-a3b-instructTakes text, vision, returns text.256K—$0.14$0.55/M——150 tok/s0.10 s
qwen3-next-80b-a3b-thinkingTakes text, vision, returns text.256K—$0.14$1.42/M——118 tok/s1.23 s
qwen3-vl-flashTakes text, vision, video, returns text.254K32K$0.02$0.21/M$0.0041/M———
qwen3-vl-flash-2026-01-22Takes text, vision, video, returns text.254K32K$0.02$0.21/M————
qwen3-max-2026-01-23Takes text, returns text.252K32K$0.45$1.80/M$0.09/M$0.56/M10 tok/s0.48 s
qwen3.6-max-previewTakes text, returns text.240K64K$1.27$7.61/M$0.13/M$1.58/M185 tok/s0.65 s
qwen3-235b-a22bTakes text, returns text.131K128K$0.28$1.12/M——81 tok/s0.63 s
bai-qwen3-vl-235b-a22b-instructTakes , returns text.131K—$0.27$1.10/M——42 tok/s1.07 s
qwen3-vl-235b-a22b-instructTakes text, vision, video, returns text.131K33K$0.27$1.10/M——30 tok/s0.47 s
qwen3-vl-235b-a22b-thinkingTakes text, vision, video, returns text.131K33K$0.27$2.74/M——58 tok/s2.22 s
qwen3-vl-30b-a3b-instructTakes text, vision, video, returns text.128K32K$0.10$0.41/M——42 tok/s1.07 s
qwen3-vl-30b-a3b-thinkingTakes text, vision, video, returns text.128K32K$0.10$1.03/M——42 tok/s1.49 s
kat-devTakes text, returns text.128K—$0.14$0.55/M————
Qwen/Qwen2.5-VL-72B-InstructTakes text, vision, video, returns text.128K—$0.50$0.50/M————
qwen3-coder-plus-2025-07-22Takes text, returns text.128K66K$0.54$2.16/M——83 tok/s1.63 s
qwen3-reranker-0.6bTakes text, vision. Output modality not published.16K8K$0.11$0.11/M————
qwen-mt-turboTakes text, returns text.16K8K$0.19$0.53/M——25 tok/s0.42 s
qwen-mt-plusTakes text, returns text.16K8K$0.49$1.48/M——69 tok/s0.41 s
qwen-audio-3.0-tts-flashTakes text, returns audio.——Free$0.14/M————
qwen-audio-3.0-tts-plusTakes text, returns audio.——Free$0.20/M————
qwen-image-3.0Takes text, vision, returns vision.——FreeFree/M————
qwen-image-3.0-proTakes text, vision, returns vision.——FreeFree/M————
qwen-flash——$0.02$0.20/M————
qwen-flash-2025-07-28——$0.02$0.20/M————
qwen-turboTakes text, returns text.——$0.05$0.09/M$0.0092/M—94 tok/s0.66 s
qwen-turbo-2024-11-01Takes text, returns text.——$0.05$0.09/M——94 tok/s0.66 s
qwen-turbo-2025-04-28Takes , returns text.——$0.05$0.09/M————
qwen-turbo-latestTakes , returns text.——$0.05$0.09/M$0.0092/M———
qwen3-0.6bTakes , returns text.——$0.05$0.46/M————
qwen3-1.7bTakes , returns text.——$0.05$0.46/M————
qwen3-4bTakes , returns text.——$0.05$0.46/M————
bce-reranker-baseTakes text, vision. Output modality not published.——$0.07$0.07/M————
qwen3-embedding-0.6bTakes text. Output modality not published.——$0.07$0.07/M————
qwen3-embedding-4bTakes text. Output modality not published.——$0.07$0.07/M————
qwen3-embedding-8bTakes text. Output modality not published.——$0.07$0.07/M————
Qwen/Qwen2-7B-Instruct——$0.08$0.08/M————
qwen3-8bTakes , returns text.——$0.08$0.80/M————
text-embedding-v4Takes text. Output modality not published.——$0.08$0.08/M————
qwen-long——$0.10$0.40/M————
qwen3-30b-a3b-instruct-2507——$0.10$0.41/M————
gte-rerank-v2Takes text, vision. Output modality not published.——$0.11$0.11/M————
qwen3-reranker-4bTakes text, vision. Output modality not published.——$0.11$0.11/M————
qwen3-reranker-8bTakes text, vision. Output modality not published.——$0.11$0.11/M————
qwen-plus——$0.11$1.13/M$0.02/M$0.14/M——
qwen-plus-2025-04-28Takes , returns text.——$0.11$1.13/M$0.02/M$0.14/M——
qwen-plus-2025-07-28——$0.11$1.13/M$0.02/M$0.14/M——
qwen-plus-latestTakes , returns text.——$0.11$1.13/M$0.02/M$0.14/M——
qwen3-30b-a3b——$0.12$1.20/M————
qwen3-30b-a3b-thinking-2507——$0.12$1.20/M————
gme-qwen2-vl-2b-instructTakes text, vision, video. Output modality not published.——$0.14$0.14/M————
Qwen/QwQ-32BTakes , returns text.——$0.14$0.56/M————
Qwen/Qwen2.5-Coder-32B-Instruct——$0.16$0.16/M————
Qwen/QwQ-32B-Preview——$0.16$0.16/M————
qwen3-14bTakes , returns text.——$0.16$1.60/M————
qwen3-32b——$0.16$0.64/M——100 tok/s1.05 s
Qwen/Qwen2-1.5B-Instruct——$0.20$0.20/M————
Qwen/Qwen3-8B——$0.20$0.20/M————
qwen2.5-coder-1.5b-instruct——$0.20$0.40/M————
qwen2.5-coder-7b-instruct——$0.20$0.40/M————
qwen2.5-math-1.5b-instruct——$0.20$0.20/M————
qwen2.5-math-7b-instruct——$0.20$0.40/M————
Qwen/Qwen2-57B-A14B-Instruct——$0.24$0.24/M————
Qwen/Qwen2.5-VL-32B-InstructTakes text, vision, video, returns text.——$0.24$0.24/M————
qwen-3-235b-a22b-instruct-2507——$0.28$1.40/M————
qwen-3-235b-a22b-thinking-2507——$0.28$2.80/M————
Qwen2-VL-7B-InstructTakes text, vision, video. Output modality not published.——$0.28$0.70/M————
Qwen3-235B-A22B-Thinking-2507——$0.28$2.80/M——87 tok/s0.27 s
qwen-max——$0.38$1.52/M————
qwen-max-0125——$0.38$1.52/M————
qwen-3-32b——$0.40$1.60/M————
qwen-qwq-32b——$0.40$0.80/M————
Qwen/Qwen2.5-7B-Instruct——$0.40$0.40/M————
Qwen/Qwen3-32B——$0.40$0.80/M————
qwen2.5-14b-instruct——$0.40$1.20/M————
qwen2.5-3b-instruct——$0.40$0.80/M————
qwen2.5-7b-instruct——$0.40$0.80/M————
Qwen/Qwen3-14B——$0.50$0.50/M————
Qwen/Qwen2.5-32B-Instruct——$0.60$0.60/M————
qwen2.5-32b-instruct——$0.60$1.20/M————
Qwen2.5-VL-72B-InstructTakes text, vision, video. Output modality not published.——$0.62$0.62/M————
Qwen/Qwen2-72B-Instruct——$0.80$0.80/M————
Qwen/Qwen2.5-72B-Instruct——$0.80$0.80/M————
Qwen/Qwen2.5-72B-Instruct-128K——$0.80$0.80/M————
qwen2.5-72b-instruct——$0.80$2.40/M————
qwen2.5-math-72b-instruct——$0.80$2.40/M————
qwen3-max-previewTakes text, vision, returns text.——$0.85$3.38/M$0.17/M—8 tok/s0.53 s
Qwen/Qwen3-30B-A3B——$1.00$1.00/M————
Qwen/QVQ-72B-Preview——$1.20$1.20/M————
qwen-imageTakes text, vision, returns vision.——$2.00$2.00/M————
qwen-image-2.0Takes text, vision, returns vision.——$2.00$2.00/M————
qwen-image-2.0-proTakes text, vision, returns vision.——$2.00$2.00/M————
qwen-image-editTakes text, vision, returns vision.——$2.00$2.00/M————
qwen-image-maxTakes text, vision, returns vision.——$2.00$2.00/M————
wan2.7-imageTakes text, vision, returns vision.——$2.00$2.00/M————
wan2.7-image-proTakes text, vision, returns vision.——$2.00$2.00/M————
Qwen2-VL-72B-InstructTakes text, vision, video. Output modality not published.——$2.18$6.54/M————
qwen2.5-vl-72b-instructTakes text, vision, returns text.——$2.40$7.20/M————
qwen-max-longcontext——$7.00$21.00/M————
wan2.2-i2v-plusTakes text, vision, returns video.——$480.00$480.00/M————
wan2.5-i2v-previewTakes text, vision, returns video.——$480.00$480.00/M————
wan2.5-t2v-previewTakes text, returns video.——$480.00$480.00/M————
happyhorse-1.0-i2vTakes text, returns video.——$720.00$720.00/M————
happyhorse-1.0-r2vTakes text, returns video.——$720.00$720.00/M————
happyhorse-1.0-t2vTakes text, returns video.——$720.00$720.00/M————
happyhorse-1.0-video-editTakes text, returns video.——$720.00$720.00/M————
happyhorse-1.1-i2vTakes text, returns video.——$720.00$720.00/M————
happyhorse-1.1-r2vTakes text, returns video.——$720.00$720.00/M————
happyhorse-1.1-t2vTakes text, returns video.——$720.00$720.00/M————
wan2.6-i2vTakes text, vision, returns video.——$720.00$720.00/M————
wan2.6-t2iTakes text, vision, returns vision.——$720.00$720.00/M————
wan2.6-t2vTakes text, returns video.——$720.00$720.00/M————
wan2.7-i2vTakes text, vision, returns video.——$720.00$720.00/M————
wan2.7-r2vTakes text, video, returns video.——$720.00$720.00/M————
wan2.7-t2vTakes text, returns video.——$720.00$720.00/M————
wan2.7-videoeditTakes text, video, returns video.——$720.00$720.00/M————

Prices are USD per million tokens; cache read and cache write are the rates for prompt-cache hits and for writing a prompt into the cache. Throughput and latency are measured on AIHubMix — the same figures the model detail page shows — not vendor claims. A dash means the catalog does not publish that field for that model, which is not the same as the model not supporting it.

Common Comparisons

Qwen/QwQ-32B vs Qwen/Qwen2.5-VL-32B-InstructQwen/QwQ-32B vs Qwen/Qwen2.5-VL-72B-InstructQwen/QwQ-32B vs bai-qwen3-vl-235b-a22b-instructQwen/QwQ-32B vs kimi-k3Qwen/QwQ-32B vs glm-5.2

Qwen on AIHubMix

Which Qwen model should I start with?

qwen3-vl-flash at $0.02/M input — the cheapest entry here that declares tool calling, and it carries a 254K context. Move up to happyhorse-1.0-i2v when answer quality matters more than cost, or to qwen3.6-plus-preview-free for long-form reasoning.

Which of these models reason before answering?

22 of the 140 models here declare a reasoning phase — they work through the problem before producing an answer, which helps on multi-step problems at the cost of extra output tokens. Use the Reasoning filter above the table to see them. The catalog does not record anything further about how they differ, so this page does not sort them into families.

Why are there several entries for the same model?

Because each row is a route you can call, not a model release. Some IDs name an upstream (azure-, alicloud-, cc-), some are the open-weight repository form (Qwen/…), and some differ only in capitalisation, kept so older integrations keep working.

The catalog does not carry a field saying which of those a given row is, so this page does not sort them into buckets it would have to invent. Every row shows that route’s own price, context and speed — compare those directly, and open a model to see the upstreams that serve it.

How is cached input billed?

The Cache read column is the rate for input tokens served from the prompt cache — for example qwen3.5-flash bills cache hits at 10% of the input rate and qwen3.5-plus bills cache hits at 10% of the input rate. Cache write is the surcharge for putting a prompt into the cache in the first place, and only a few upstreams bill it separately. A dash in either column means the catalog carries no cache rate for that model, so plan on paying the full input rate.

Do I need a separate Qwen account?

No. One AIHubMix key covers every model on this page, and switching between them is a change to the model string — billing, rate limits, and logs stay in one place.

Other Providers

OpenAI
132 modelsGoogle
81 modelsZ.AI
64 modelsLlama
52 modelsByteDance
46 models

Start calling Qwen in one line

One key, one endpoint, 844 models across 35 providers.

Get API keyRead the docs
AIHubMix

© 2023 - 2026 AIHubMix, LLC

Product

Company

Developer

Connect