Kyma APIKyma API
Alibaba

Alibaba

Qwen 3.7 Flash

Qwen 3.7 Flash is Alibaba’s vision-language reasoning model built for high-volume multimodal tasks. Reach for it when you need fast, cheap-tier processing of text, images, and video with tool use and reasoning capabilities.

Modalities

Text+Image → Text

Below release

Not established

Weights

Not established

Where it sits in the catalogue Kyma measures

Every number against every text model Kyma prices per token — a stated rule, not a chosen line-up.

Price · input#1/84

$0.0435

output $0.1886 · cached $0.0081

median $1.013 · best $0.0435

Price · output#1/83

$0.1886

per 1M tokens generated

median $3.528 · best $0.1886

Availability · 30d#38/54

99.81%

537 observations since 2026-07-29

median 100.0% · best 100.0%

Throughput

90 tok/s

from traffic, 7d — real prompts, not a probe

median 46.1 tok/s · best 160.1 tok/s

Response time

31.76 s

from traffic, 7d — real prompts, not a probe

median 1.33 s · best 0.46 s

Context

1M

max output 8,192

published by Alibaba, not measured here

Tick above each rail is this model, below it the other 83. Dashed rule is the field median, solid is its best. Better is left; the axis stops at the 90th percentile, so a few models sit past its right edge.

Usage

How much this model is actually called here.

Rank

#15

of 113 active models

Tokens served

63.6M

all-time

Platform share

1.4%

of all tokens

Tokens · last 15 daysOct 15 → Oct 29

How it behaves under real clients

The same model answering different prompt shapes — measured, not benchmarked.

ClientRequestsTokensThroughputTo first tokenp95 totalCompleted
OpenAI Python1161.1M58 tok/s3.32 s32 s100%
OpenCode107497.0K82 tok/s1.77 s23 s100%

Grouped by the client that sent the request. Not a benchmark: the same model on different prompt shapes. Throughput and time-to-first-token are medians; p95 is the slowest response in twenty. 3 clients under 100 requests not shown.

Pricing

Pay per token. Cached input bills at this model’s own cached rate, listed below.

$0.0435 /1M input$0.1886 /1M output
Hobby10 req/day · 2K in / 500 out
~$0.05/mo
Production1,000 req/day · 2K in / 500 out
~$5.44/mo
Scale20,000 req/day · 2K in / 500 out
~$109/mo
+ Estimate your workload
1,000
2,000
500
30%

Estimated monthly cost

$4.80

$0.1601 / day on Qwen 3.7 Flash

Same workload on:

Qwen 3.8 27B$63.94+1232%
Qwen 3.8 Max$234+4772%

Estimates use list pricing with cached input at this model's own cached rate. Actual bills depend on real token counts, and every response includes its exact cost.

When to use Qwen 3.7 Flash

Updated 2026-07-29

Where this model earns its cost — and where it doesn't.

This is a vision-language reasoning model from Alibaba that accepts text, image, and video inputs and returns text. It supports a 1M token context window and outputs up to 8,192 tokens. The model operates in the cheap cost tier and is optimized for high-throughput workloads.

On Kyma, it runs through an OpenAI-compatible endpoint with automatic request failover if a serving path degrades. Prompt caching is enabled, billing repeated prefixes at this model's cached input rate. Responses include the exact request cost in usage.cost and the executed model ID in the X-Kyma-Model header.

It does not support structured outputs, so you will need to parse JSON or enforce formatting manually. The 1M context window is large, but the 8,192 token output limit restricts long-form generation.

Multimodal agent orchestration

Route text, image, and video inputs to tool-calling agents that run at high volume.

Visual code review

Analyze screenshots, UI mockups, and video walkthroughs to generate code suggestions.

Document analysis at scale

Process large batches of mixed-media files within a 1M token context window.

High-throughput chat routing

Handle conversational workloads with fast response times and low compute overhead.

Not ideal for: Avoid this model when you require guaranteed structured JSON outputs or need to generate responses longer than 8,192 tokens.

How it compares

Against the peers people actually weigh it against.

SpecQwen 3.7 FlashQwen 3.8 27BQwen 3.8 Max
Input /1M$0.0435$0.3826$2.228
Output /1M$0.1886$2.732$6.684
Context1M1M1M
ToolsYesYesYes
ReasoningYesYesYes
Throughput90 tok/s

Quick start

Up and running in under two minutes.

  1. 1

    Create an API key

    Sign up and grab a key from the dashboard — $0.50 free credit on the free tier, which covers Qwen 3.7 Flash. No card required.

    Get API key →
  2. 2

    Make your first request

    Drop in your key and send a chat completion — fully OpenAI-compatible.

    curl https://kymaapi.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.7-flash",
    "messages": [
      {"role": "user", "content": "Explain prompt caching in one paragraph."}
    ]
  }'
  3. 3

    Stream responses

    Add "stream": true to receive tokens as they arrive.

    curl https://kymaapi.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.7-flash",
    "stream": true,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

FAQ

Common questions about this model.

What is the context window of Qwen 3.7 Flash?

Qwen 3.7 Flash has a 1M-token context window — roughly 1471 pages of text in a single request.

How much does the Qwen 3.7 Flash API cost?

$0.0435 per 1M input tokens and $0.1886 per 1M output tokens, with cached input at $0.0081/1M — 81% off the input rate on repeated prompt prefixes. No subscription; you pay only for what you use.

Does Qwen 3.7 Flash support function calling?

Yes — Qwen 3.7 Flash supports tool/function calling and structured outputs (JSON mode), so it works with agent frameworks out of the box.

Are the weights for Qwen 3.7 Flash publicly available?

Kyma cannot say. We could not establish it from Alibaba's own material — neither a published checkpoint under Alibaba's own account, nor a statement from Alibaba that the weights are withheld. So we publish "not established" and make no claim in either direction. Treat that as Kyma declining to guess, not as evidence about Alibaba: the honest answer to a question nobody sourced is that nobody sourced it.

Is Qwen 3.7 Flash ever served below the precision its creator released it at?

Kyma cannot say, and the reason runs one step deeper: we could not even establish whether Alibaba publishes Qwen 3.7 Flash's weights, so there may be no released checkpoint to compare a route against at all. We publish "not established" rather than reading a low format as a downgrade: on several models in this catalogue a low format is exactly how the creator ships it.

How do I use Qwen 3.7 Flash?

Kyma is OpenAI-compatible: point your SDK's base URL at https://kymaapi.com/v1, use your Kyma API key, and set the model to qwen3.7-flash. Signing up is free and includes $0.50 of credit on the free tier, which covers this model — no card required.

Does this model support function calling?

Yes, it supports tool use and reasoning, allowing you to pass tool definitions and parse the model's calls.

How does prompt caching affect billing?

Kyma bills cached prompt prefixes at this model's cached input rate, reducing costs for repeated system prompts or conversation history.

What happens if the serving path fails?

Kyma automatically reroutes your request to an alternate serving path, ensuring high availability without manual retries.

Start with $0.50 free credit on the free tier — no card required.Create account →

More models by Alibaba

See all 14 →
ModelContextInputOutput
AlibabaQwen 3.8 Flash1M$0.2045$0.641
AlibabaQwen 3.8 27B1M$0.3826$2.732
AlibabaQwen 3.8 Max1M$2.228$6.684
AlibabaQwen 3.7 Plus1M$0.4325$1.731
AlibabaQwen 3.7 Max1M$2.071$6.207
AlibabaQwen 3.6 Plus1M$0.4388$2.633
AlibabaQwen 3 Coder131K$0.3546$1.612
AlibabaQwen3 Reranker 8B41K$0.135$0.00