Qwen3.7-Max was Alibaba's flagship before Qwen3.8-Max and is still sold on Model Studio. It costs $2.50 per 1M input tokens and $7.50 per 1M output tokens, with a 1M context window.
Qwen3.7-Max costs $2.50 input / $7.50 output per 1M tokens. Its successor Qwen3.8-Max is cheaper ($2.00 / $6.00), so Qwen3.7-Max mainly makes sense for teams that have already validated prompts against a pinned 3.7 snapshot.
| Qwen3.7-Max | |
|---|---|
| Provider | Alibaba |
| Tier | Mid-tier |
| Input / 1M tokens | $2.50 |
| Output / 1M tokens | $7.50 |
| Cache hit / 1M | $0.25 |
| Cache write / 1M | — |
| Context window | 1M |
| Notes | Prior Qwen flagship; strong multilingual |
| 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-Max | $80.00 | $81.00 | $3.02 |
| Mistral Large 3 | $16.00 | $16.20 | $2.23 |
| DeepSeek V4-Pro | $21.12 | $21.38 | $0.64 |
| Mistral Large 4 | $22.40 | $22.68 | $0.85 |
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 | 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 | Tier | Input / 1M | Output / 1M | Cache hit / 1M | Context | Notes |
|---|---|---|---|---|---|---|
| Mistral Large 3 | Mid-tier | $0.50 | $1.50 | — | 256K | Former Mistral flagship; deprecated Oct 31, 2026, retiring Apr 30, 2027 — replaced by Mistral Large 4 |
| DeepSeek V4-Pro | Mid-tier | $0.66 | $1.98 | $0.022 | 1M | A planned Sep 14, 2026 reroute to V4.1-Flash rates was reversed after user pushback; billing unchanged since the Aug 13, 2026 GA build. Off-peak rate shown; peak (01:00-04:00 & 06:00-10:00 UTC weekdays) is 2x: $1.32/$3.96 |
| Mistral Large 4 | Mid-tier | $0.70 | $2.10 | $0.07 | 1M | New Mistral flagship (Oct 6, 2026; public preview); open-weight 1.05T MoE (52B active), multimodal; replaces Large 3 and Medium 3.5 |
| Grok 4.3 | Mid-tier | $1.25 | $2.50 | $0.20 | 1M | Reasoning workhorse; replaced Grok 4; ≥200K tokens: $2.50/$5, cache $0.40 |
| Gemini 3.8 Flash | Mid-tier | $0.75 | $3.75 | $0.075 | 1M | On the Gemini API free and paid tiers; Gemini 3.7 Flash retired Oct 8, 2026 and now auto-routes here (requests still billed at this rate); intro price thru Dec 31, reverts to $1.50/$7.50 (cache $0.15) Jan 1 |
Alibaba sells several Qwen3.7-Max IDs at the same price. The alias qwen3.7-max currently points to the qwen3.7-max-2026-05-20 snapshot and supports both thinking and non-thinking modes; qwen3.7-max-2026-06-08 is a later snapshot; and qwen3.7-max-preview is thinking-only. Only the alias carries the 1M-token free quota in the Singapore region. Pin a dated snapshot in production so the alias moving doesn't change your outputs or token counts without warning.
Unlike Qwen3.8-Max, Qwen3.7-Max follows Model Studio's standard cache pricing: implicit (automatic) cache hits are billed at 20% of the input rate, and explicit cache hits at 10%, which is the $0.25 shown in the table. Creating an explicit cache costs 125% of the normal input price, so it pays off once a cached prefix is reused more than once inside its five-minute lifetime (each hit resets the timer). Prompts under 1,024 tokens aren't cached.
On price alone, yes: Qwen3.8-Max costs less on both input and output and offers the same 1M context. The trade-offs are behavioral and operational. Re-run your evaluation set before switching, since prompts tuned for one model don't always transfer, and note that Qwen3.8-Max's cache-hit pricing differs from the 20%/10% rule, so recheck the economics if your workload relies heavily on caching.
$2.50 for input and $7.50 for output, with explicit cache hits at $0.25.
Both cost the same. The qwen3.7-max alias supports thinking and non-thinking modes; qwen3.7-max-preview supports thinking mode only.
Yes. Alibaba still lists Qwen3.7-Max and its dated snapshots on Model Studio, alongside the newer and cheaper Qwen3.8-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.