Three tiers
Max released May 19, 2026, Plus on June 1, 2026 and Flash on July 27, 2026. Pick per task, one balance covers all of them.
Qwen3.7 is Alibaba's current generation, and it ships in three tiers: Max for the hardest reasoning, Plus for everyday work, and Flash, released on July 27, 2026, a vision language model that costs 0.03 USD per million input tokens. All three carry a 1 million token context window, and Flash and Plus read images. On avots.ai you run them pay per use, with no Alibaba Cloud account, and the same balance also covers Claude Opus 5, GPT-5.6, Gemini and 40+ image, video and music models.
Max released May 19, 2026, Plus on June 1, 2026 and Flash on July 27, 2026. Pick per task, one balance covers all of them.
Every tier takes a full million tokens in one conversation, with output limits of 128K on Max and Plus.
Flash input costs 0.03 USD and output 0.13 USD per million tokens, which is among the cheapest vision capable models anywhere.
On the independent Artificial Analysis Intelligence Index, well above the median for reasoning models in its price tier.
Max generates at roughly 203 tokens per second in Artificial Analysis testing, fast for a flagship reasoning model.
Flash takes text, images and video, Plus takes text and images. Max is text only, so pair it with Flash when the task starts from a screenshot.
| Qwen3.7 Max | Qwen3.7 Plus | Qwen3.7 Flash | |
|---|---|---|---|
| Price per 1M in / out | $1.475 / $4.425 | $0.32 / $1.28 | $0.03 / $0.13 |
| Intelligence Index | 46 | 39 | not published |
| Output speed | about 203 tokens/s | about 53 tokens/s | not published |
| Context window | 1M tokens | 1M tokens | 1M tokens |
| Max output | 128K tokens | 128K tokens | 64K tokens |
| Image input | no | yes | yes, plus video |
| On avots.ai | pay per use | pay per use | pay per use |
Max costs about 45 times more per input token than Flash, and it buys you the higher index score and roughly four times the generation speed. For most everyday work Plus or Flash is the honest answer, and switching between them in the picker takes one click.
A native vision language model tuned for object recognition, spatial understanding and real world perception, not a text model with an adapter.
Alibaba positions it for multimodal agents, visual coding, search and computer interaction, with tool calling built in.
It accepts video alongside text and images, which almost nothing else at this price does.
At 0.03 USD per million input tokens you can push a lot of screenshots through it before the cost is worth thinking about.
Sign up on app.avots.ai with email, Google or Telegram. Takes a minute.
1,000 tokens cost 0.99 EUR. No plan to choose, no renewal, the balance sits there until you use it.
Choose Max, Plus or Flash in the model picker, or let Agent Avots route for you. Every reply shows its token cost.
app.avots.ai on desktop and mobile, no install needed.
Qwen3.7 in @AvotsAIbot, same balance, from any phone.
Use it inside Claude Desktop, Cursor or Cline via the MCP server.
The OpenAI compatible API takes your existing client, one key for every model.
1,000 tokens cost 0.99 EUR and a Flash answer costs a handful of them. The same balance unlocks Claude Opus 5, GPT-5.6, Gemini, Veo video and Nano Banana images, with no Alibaba Cloud account.
Sign up and chat ✦Newer version. This model has a successor — same balance, same app.
The previous Alibaba generation, still on the same balance.
Every LLM on avots.ai and which one to pick per task.
284B mixture of experts built for agent loops, 1M context.
Moonshot's frontier open model, number one at coding, 1M context.
Anthropic's newest flagship, 1M context, 96% SWE-bench Verified.
What to use instead of ChatGPT Plus and how one balance replaces it.
avots.ai runs 40+ AI models for chat, image, video and audio on one balance. See the full list on the Tools page or open the app and start creating.