Qwen3.5 Flash Has Been Retired — Here Are the Cheapest Capable LLM APIs in 2026

Published 2026-07-20 · Updated 2026-09-16 · NovAI Team

⚠️ Model retirement notice (September 2026): Alibaba has discontinued Qwen3.5 Flash and Qwen3.5 Plus upstream, and both have been removed from the NovAI platform. Calls to model id qwen3.5-flash or qwen3.5-plus will now fail. This guide has been updated to point you to the best low-cost replacements, which you can switch to by changing a single model parameter.

Cost is still the top constraint for high-volume pipelines — and the budget tier on NovAI is as strong as ever. With Qwen3.5 Flash gone, the cheapest capable models are now DeepSeek V4 Flash ($0.08 / $0.17 per 1M), the new GLM-5.3-Flash ($0.08 / $0.28, 1M context), and Doubao-Seed-2.0-Lite ($0.075 input). Here's how they compare and when to use each.

Cheapest models on NovAI (per 1M tokens)

Model	Input	Output	Context
doubao-seed-2.0-lite	$0.075	$0.45	256K
deepseek-v4-flash (best value)	$0.08	$0.17	1M
glm-5.3-flash (NEW)	$0.08	$0.28	1M
hy3-preview (Hunyuan)	$0.127	$0.423	256K

At $0.075–$0.08 per 1M input tokens, these are among the lowest-cost capable models available anywhere — and DeepSeek V4 Flash's $0.17 output price makes it the best all-round value for generation-heavy workloads.

Best use cases

High-volume classification, tagging, extraction, and routing.
Chatbots and summarization at scale.
Cost-sensitive background jobs where you don't need a flagship model.

Quickstart (migrating from qwen3.5-flash)

from openai import OpenAI

client = OpenAI(
    api_key="nvai-your-key",
    base_url="https://aiapi-pro.com/v1"
)

resp = client.chat.completions.create(
    model="deepseek-v4-flash",   # was "qwen3.5-flash"; or use "glm-5.3-flash"
    messages=[{"role": "user", "content": "Hello!"}]
)
print(resp.choices[0].message.content)

Ship at $0.08/1M input with DeepSeek V4 Flash or GLM-5.3-Flash — plus $2 free credit to start.

Get $2 Free Credit — No Credit Card →

Frequently Asked Questions

Is Qwen3.5 Flash still available on NovAI?

No. Qwen3.5 Flash and Qwen3.5 Plus were retired upstream and removed from NovAI in September 2026. Calls to model id qwen3.5-flash or qwen3.5-plus now fail. Switch to DeepSeek V4 Flash, GLM-5.3-Flash, or Doubao-Seed-2.0-Lite.

What is the cheapest LLM API in 2026 now?

Doubao-Seed-2.0-Lite has the cheapest input at $0.075 per 1M tokens. DeepSeek V4 Flash ($0.08 input / $0.17 output) is the best overall value with the lowest output price, and the new GLM-5.3-Flash ($0.08 / $0.28) adds a 1M-token context window — all available on NovAI.

What should I migrate my qwen3.5-flash workload to?

For high-volume classification, extraction, summarization, and routing, migrate to deepseek-v4-flash or glm-5.3-flash — both are 1M-context budget models at $0.08 per 1M input tokens. Just change the model id in your existing OpenAI-compatible calls. For complex multi-step reasoning, step up to qwen3.8-max or DeepSeek V4-Pro.