QWEN-PLUS

Qwen-Plus: Alibaba's balanced mid-tier

Strong performance-to-cost ratio. 128K context. Thinking and non-thinking modes. Native web search. $0.42 input / $1.26 output per million tokens.

Try Qwen-Plus on Council See it in a council

Qwen-Plus is Alibaba's balanced mid-tier model — the workhorse below the Qwen3-Max and Qwen3.6-Max flagships. It ships a 128K context window, native web search via DashScope, and supports both thinking and non-thinking modes (flip the deeper reasoning chain on only when you need it). Priced at $0.42 per million input tokens and $1.26 per million output, it sits between DeepSeek V4 Flash ($0.14/$0.28) and the Qwen Max tier ($1.04+/$6+) — roughly 12x cheaper than GPT-5.5 on input. The case for: everyday workloads where you don't need frontier-tier quality but want better-than-cheap behavior and the option to flip thinking mode on. Great for summarization, classification, light coding, customer-facing chat, and RAG pipelines. The case against: hard reasoning, complex refactors, and tasks demanding the absolute coding benchmark ceiling — for those, step up to Qwen3.6-Max or Opus 4.7.

Specs

ProviderAlibaba (Qwen)
Context window131,072 tokens
ReasoningThinking + non-thinking modes
Web searchNative (DashScope enable_search)
Price (in / out per 1M)$0.42 / $1.26
MultimodalText only
TierBalanced mid-tier

Best at

Where it loses

Frequently asked questions

How does Qwen-Plus compare to Qwen3-Max?

Qwen-Plus is the balanced mid-tier; Qwen3-Max is the trillion-parameter flagship. Plus is roughly 3x cheaper on input and 5x cheaper on output. Pick Plus for everyday work, Max for hard reasoning.

What are thinking vs non-thinking modes?

Non-thinking is the fast default — direct answers. Thinking mode produces an internal chain-of-thought before responding, better on hard reasoning at the cost of latency and tokens.

How much does Qwen-Plus cost?

$0.42 per million input tokens and $1.26 per million output. Roughly 12x cheaper than GPT-5.5.

Does Qwen-Plus support web search?

Yes — natively via DashScope's enable_search flag, the same mechanism Council AI uses for Qwen-Plus, Qwen3-Max, and Qwen3.6-Max.

What's the qwen-plus-latest naming about?

Alibaba ships rolling updates to Qwen-Plus and exposes them under the qwen-plus-latest endpoint alias. You always get the most recent stable snapshot.

Is Qwen-Plus good for production chat?

Yes — it's tuned for everyday workloads with strong latency, native search, and a 128K window. A common pick for customer-facing chat at scale.