model comparison

vs

Qwen3.5 397B vs Qwen3.7 Plus

two Qwen models, compared on listed price, real workload cost, context and capabilities.

the short answer

  • Qwen3.7 Plus lists lower input and output prices.
  • Qwen3.7 Plus supports vision; Qwen3.5 397B does not.
  • Qwen3.7 Plus reads cached input for less ($0.08/M against Qwen3.5 397B's $0.35/M), which decides the bill for agents that resend long context.

side-by-side specs

specQwen3.5 397BQwen3.7 Plus
providerQwenQwen
input price$0.6/M$0.4/M
output price$3.6/M$1.6/M
cached input$0.35/M$0.08/M
blended price (3:1)$1.35/M$0.7/M
context window256k tokens256k tokens
max output—131k tokens
inputtexttext, images, video
structured outputyesyes
weightsopen (Apache 2.0)closed
releasedFebruary 2026—

same work, both models

workloadQwen3.5 397BQwen3.7 Plus
one chat reply3,000 tokens in, 500 outQwen3.5 397B costs 1.8× as much as Qwen3.7 Plus$0.0036$0.002
summarise a 100-page report50,000 tokens in, 1,500 out, one requestQwen3.5 397B costs 1.6× as much as Qwen3.7 Plus$0.04$0.02
1,000 classification calls500 tokens in, 20 out, per call$0.37$0.23
coding-agent session2M tokens in across ~60 turns, 60,000 out, no cache$1.42$0.90
same session, cached85% of input read from the prompt cache, the rest written to itQwen3.5 397B costs 2.8× as much as Qwen3.7 Plus$0.99$0.35

at listed rates. token counts are typical sizes, not measurements.

when to pick which

Qwen3.5 397B

Qwen3.5 397B is the open-weight Qwen3.5-397B-A17B model from Qwen, with 397B total and 17B activated parameters, released in February 2026. the weights accept images and video, but this face is served text-only.

pick it if

  • want Apache 2.0 open weights you could later self-host
  • want a thinking-by-default model with tool calling
  • reuse long prompts: cache hits bill at $0.35 versus $0.60 input

skip it if

  • you need image input: Qwen3.7 Plus takes images and costs less
  • your workload is output-heavy and a cheaper sibling covers it

Qwen3.7 Plus

Qwen3.7 Plus is Alibaba's multimodal agent model from the Qwen3.7 generation, served through Model Studio with text, image and video input. in this catalog it sits between Qwen3.8 Omni Flash and Qwen3.8 Max on price.

pick it if

  • you want a Qwen model with image and video input below Qwen3.8 Max pricing
  • you reuse long prompts: cache hits bill at a fifth of input
  • you run screen, GUI or tool-using agents, the work Qwen pitches it for

skip it if

  • you need audio input; Qwen3.8 Omni Flash takes it for less
  • you need more than the 256K context window this catalog serves

faq

which is cheaper, Qwen3.5 397B or Qwen3.7 Plus?

Qwen3.7 Plus lists lower input and output prices. on a coding-agent session (2M tokens in, 60,000 out) Qwen3.5 397B costs $1.42 and Qwen3.7 Plus $0.90.

which one accepts images?

Qwen3.7 Plus supports vision; Qwen3.5 397B does not.

can i switch between Qwen3.5 397B and Qwen3.7 Plus mid-task?

yes. in an octomind session, /model octohub:qwen3.5-397b and /model octohub:qwen3.7-plus switch models without losing the conversation, so you can draft on one and review on the other.

try either model

switch between models inside the same octomind session.

prices from the catalog of 2026-11-18