Qwen3.7-Plus

What it is. Qwen3.7-Plus (equivalent to qwen3.7-plus-2026-05-26) is Alibaba Cloud’s previous-generation plus-tier API model: a closed-weight, proprietary mid-tier SKU on Model Studio. Why it matters. It is a documented, billable mid-tier option that remains listed after Alibaba’s flagship line moved to Qwen3.8 — a live example of how Alibaba tiers its lineup by price while letting older SKUs keep billing for pinned integrations. Key characteristics. $0.4 input / $1.6 output per 1M tokens (≤256K tier) on the International region, a higher 256K–1M tier at $1.2 / $4.8, and a limited-time 20% console discount. What a professional should know. It has been superseded in the flagship line by Qwen3.8-Max, and its context window, capabilities and benchmark results are not published on the official pricing page as of 2026-09-27.

Status and successor

Qwen3.7-Plus is superseded in Alibaba’s line-up by Qwen3.8-Max but is not discontinued — it remains an active, billable SKU on Model Studio. This matters for existing integrations: pinned deployments on qwen3.7-plus keep working and billing, but new workloads should evaluate Qwen3.8-Max first. China AI Hub analysis indicates this superseded-but-active state is the standard Chinese API pattern for mid-tier models — vendors rarely hard-deprecate a plus-tier SKU, they simply stop marketing it and let price-tier positioning do the migration.

What is and is not documented

The pricing page is the only substantive source: it documents price tiers but does not publish a context window, capability list or benchmark results for Qwen3.7-Plus. China AI Hub analysis indicates this thin documentation is itself the clearest signal of the model’s position — a superseded mid-tier SKU is documented only for billing, not for evaluation, which is why the database records it as partially_verified with the capability gap explicitly noted. The two context-tier price bands (≤256K and 256K–1M) imply a 1M-context option exists, but that is an inference from pricing, not a published specification.

Pricing detail

List price is $0.4 / $1.6 (≤256K) and $1.2 / $4.8 (256K–1M) on the International region; the Beijing region lists $0.276 / $1.101 and $0.826 / $3.301. The 3x jump between the ≤256K and 256K–1M tiers is the operative cost fact for long-context use. The two context tiers imply the model supports up to a 1M context option, though the exact context window is not published on the official pricing page as of 2026-09-27 — recorded here as not publicly documented. The limited-time 20% console discount is a vendor-stated promotion, not a permanent list price.

How it relates to Qwen3.8

China AI Hub analysis indicates Qwen3.7-Plus occupies the previous-generation mid-tier: it is cheaper than Qwen3.8-Max ($0.4/$1.6 vs $1.65–$2/$4.951–$6) but lacks a published benchmark or capability card. The successor publishes a full architecture card (2.4T MoE, 95B active, Gated DeltaNet hybrid) and five vendor benchmark rows, whereas Qwen3.7-Plus publishes neither. For cost-sensitive text workloads where a mid-tier plus model is sufficient, it remains a documented, billable option; for capability-critical work, the successor’s published benchmarks argue for Qwen3.8-Max. See the state of China’s AI models research.

Suitable and less suitable workloads

Well-suited: cost-sensitive text workloads where a mid-tier plus model is sufficient, and pinned integrations that already call qwen3.7-plus and prefer billing continuity over migration. Less suited: capability-critical work (no published benchmarks), long-context work above the ≤256K tier (3x price jump), and any deployment that needs a documented context window, capability list or self-hosting path (closed-weight, proprietary).

Where this model fits

Workload	Relevance
Long-context analysis	Unknown
Coding	Unknown
Structured API workflows	Moderate
Agent orchestration	Unknown
Local self-hosted deployment	No evidence
GUI automation	No evidence
Video generation	No evidence
Enterprise cloud	Moderate

Relevance judgments are China AI Hub analysis based on documented capabilities, not vendor claims. Capability-specific rows are Unknown because Alibaba publishes no capability or benchmark data for this model.

China AI Hub analysis

China AI Hub analysis indicates Qwen3.7-Plus is best read as a legacy billing SKU rather than a model Alibaba wants you to evaluate: it is documented only for price, superseded in the flagship line by Qwen3.8-Max, and left active to avoid breaking pinned integrations. Its presence in the database is useful precisely as a cautionary data point — a model that is billable but effectively undocumented on capabilities illustrates how much of the Chinese API layer is priced-positioned rather than specification-positioned, and why “active on the pricing page” is not the same as “recommended for new work.”

What is uncertain

• The context window, capabilities and benchmark results are not published on the official pricing page as of 2026-09-27.
• Whether a 1M-context option exists is inferred from the two pricing tiers, not confirmed by a published specification.
• The limited-time 20% discount’s end date is not documented.

Sources

evidence_id	source_name	source_url	source_type	published	verified	confidence	conflict
src-models-qwen3.7-plus-1	Alibaba Cloud Model Studio — model pricing	https://www.alibabacloud.com/help/en/model-studio/model-pricing	Official documentation	—	2026-09-27	high	—

Labels used above: Official fact (pricing and SKU facts from Model Studio), Vendor-reported claim (the 20% limited-time discount, as stated by Alibaba), and China AI Hub analysis (our synthesis, introduced as such). Context window, capabilities and benchmark scores are not publicly documented for this model.