Alibaba sells Qwen models through Model Studio. The rates below are its international (Singapore) list prices for the three Qwen text models tracked here — Qwen3.8-Max, the older Qwen3.7-Max and the budget Qwen3.7-Plus — followed by workload costs and the cache and batch rules that move the final bill.
Qwen3.8-Max is $2.00 input / $6.00 output per 1M tokens, cheaper than the previous flagship Qwen3.7-Max ($2.50 / $7.50) with the same 1M window. For cost-sensitive traffic, Qwen3.7-Plus starts at $0.40 / $1.60, rising to $1.20/$4.80 once a request's input passes 256K tokens.
| Model | Tier | Input / 1M | Output / 1M | Cache hit / 1M | Context | Notes |
|---|---|---|---|---|---|---|
| Qwen3.8-Max | Flagship | $2.00 | $6.00 | $0.17 | 1M | New Alibaba flagship (Aug 3, 2026); 2.4T MoE |
| Qwen3.7-Max | Mid-tier | $2.50 | $7.50 | $0.25 | 1M | Prior Qwen flagship; strong multilingual |
| Qwen3.7-Plus | Budget | $0.40 | $1.60 | $0.04 | 1M | Balanced Qwen workhorse; base tier list price shown (0-256K tokens); 256K-1M tokens: $1.20/$4.80; both bands currently carry a limited-time 20% discount |
| Model | Support chatbot 10,000 conversations × (2,000 input + 400 output tokens) | Document Q&A 1,000 questions × (30,000 input + 800 output tokens) | Coding agent session 4M input tokens (90% cache hits) + 150,000 output tokens |
|---|---|---|---|
| Qwen3.7-Plus | $14.40 | $13.28 | $0.54 |
| Qwen3.8-Max | $64.00 | $64.80 | $2.31 |
| Qwen3.7-Max | $80.00 | $81.00 | $3.02 |
Standard rates per 1M tokens from data on this site. The coding-agent row assumes 90% of input is read from the prompt cache (full input price where a model has no cache rate). Cache-write fees, batch discounts and long-context surcharges listed in the notes are not included. Updated November 2, 2026.
| Model | Provider | Tier | Input / 1M | Output / 1M | Context |
|---|---|---|---|---|---|
| Qwen3.8-Max | Alibaba | Flagship | $2.00 | $6.00 | 1M |
| Grok 4.7 | SpaceXAI | Flagship | $2.00 | $6.00 | 500K |
| Gemini 4 Argon | Flagship | $2.00 | $10.00 | 1M | |
| Kimi K3.1 Pro | Moonshot AI | Flagship | $2.00 | $10.00 | 1M |
| Gemini 3.1 Pro | Flagship | $2.00 | $12.00 | 1M | |
| Qwen3.7-Max | Alibaba | Mid-tier | $2.50 | $7.50 | 1M |
| Mistral Large 3 | Mistral | Mid-tier | $0.50 | $1.50 | 256K |
| DeepSeek V4-Pro | DeepSeek | Mid-tier | $0.66 | $1.98 | 1M |
| Mistral Large 4 | Mistral | Mid-tier | $0.70 | $2.10 | 1M |
| Grok 4.3 | SpaceXAI | Mid-tier | $1.25 | $2.50 | 1M |
| Qwen3.7-Plus | Alibaba | Budget | $0.40 | $1.60 | 1M |
| Ministral 3 14B | Mistral | Budget | $0.20 | $0.20 | 256K |
| Gemini 2.5 Flash-Lite | Budget | $0.10 | $0.40 | 1M | |
| GPT-6 Luna | OpenAI | Budget | $0.10 | $0.50 | 1M |
| Claude Haiku 5.5 | Anthropic | Budget | $0.10 | $0.50 | 1M |
Alibaba's cheapest model per tier (bold) against the four cheapest same-tier models from other providers, by output price.
Qwen3.7-Plus has two price bands, and Alibaba bills the whole request at the band that matches its total input — not just the tokens above the threshold. A 300K-token prompt is therefore billed entirely at $1.20 input / $4.80 output per 1M, three times the base $0.40 / $1.60. For long-document work, trim or chunk prompts to stay under 256K, or price the job on Qwen3.8-Max, which has a single rate all the way to 1M input tokens. Alibaba marks both Qwen3.7-Plus bands as a limited-time 20% discount, so expect them to rise when it ends.
Model Studio has two cache modes. Explicit caching, where you mark the prefix yourself, bills cache creation at 125% of the normal input price and gives the cheapest hits; the cache-hit prices in the table match that explicit rate (10% of input for the 3.7 models). Implicit caching is automatic and needs no code changes, but a hit costs 20% of input on Qwen3.7-Max and Qwen3.7-Plus, and Alibaba says Qwen3.8-Max uses its own cache ratios, shown in the console. Both modes need a prefix of at least 1,024 tokens; explicit entries live for 5 minutes and reset on each hit, so explicit caching pays off for agents and chat sessions that reuse a long prefix quickly.
Batch calls cost 50% of real-time prices and can't be combined with cache discounts. On the Singapore endpoint, batch covers Qwen3.7-Plus and Qwen3.8-Max — the latter currently for multimodal input only, not text-only requests — while Qwen3.7-Max batch is listed only in the China (Beijing) region. Each model also comes with a 1M-token free quota in Singapore, valid for 90 days; dated snapshots such as qwen3.7-plus-2026-05-26 carry no free quota.
On list price, no: Qwen3.8-Max is cheaper on both input and output, so moving over cuts the bill on every workload above. If you need a frozen version for evaluation or compliance, pin the dated snapshot qwen3.8-max-0902, which is listed at the same price as qwen3.8-max. Against Western flagships, Qwen3.8-Max's $6.00 output rate undercuts Gemini 4 Argon ($10.00) and Claude Opus 5.5 ($20.00).
On Alibaba Model Studio (Singapore), Qwen3.8-Max is $2.00 input / $6.00 output per 1M tokens, Qwen3.7-Max is $2.50 / $7.50, and Qwen3.7-Plus is $0.40 / $1.60 for requests up to 256K input tokens.
Requests with more than 256K input tokens are billed entirely at the higher band ($1.20 input / $4.80 output per 1M), not only the portion above 256K.
Yes. Batch calls cost 50% of the real-time price, but the discount can't be stacked with cache discounts, and model availability for batch differs between the Singapore and Beijing regions.
New Singapore-region users get a 1M-token free quota per model, valid for 90 days from activation or model release. Dated snapshot models are excluded.
Yes. Qwen3.8-Max lists at $2.00 / $6.00 per 1M tokens versus $2.50 / $7.50 for Qwen3.7-Max.
How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.