claude-fable-5-2 vs deepseek-v4-pro: pricing and capabilities compared
In each model's cheapest group on this site, claude-fable-5-2 costs ¥32 input and ¥160 output (Claude Plus group) and deepseek-v4-pro costs ¥16.8 input and ¥33.6 output (China Models group), per 1M tokens.
deepseek-v4-pro is cheaper on input, output and cache reads, by 48%, 79% and 30% respectively.
claude-fable-5-2 is offered in 4 groups and deepseek-v4-pro in 2 groups, with no group in common: give one key several groups (up to 5, in priority order) or use two keys.
Prices updated Nov 13, 2026, 6:27 PM
At a glance
| Item | claude-fable-5-2 | deepseek-v4-pro |
|---|---|---|
| Vendor | Anthropic | DeepSeek |
| Billing | Per token | Per token |
| Cheapest group | Claude Plus | China Models |
| Input ¥ per 1M tokens | ¥32 | ¥16.848% lower |
| Output ¥ per 1M tokens | ¥160 | ¥33.679% lower |
| Cache read ¥ per 1M tokens | ¥0.8 | ¥0.56130% lower |
| Groups | 4 groups | 2 groups |
| Input types | text and images | — |
| Output types | text | — |
| Capability tags | Vision, Multimodal, Reasoning and File analysis | Vision |
| Endpoints | Chat Completions and Claude Messages | Chat Completions |
| Available since | Nov 10, 2026 | Aug 22, 2026 |
Prices are each model's cheapest group, in CNY; the side marked "lower" is cheaper. Data comes from our price list and model catalog; anything not listed shows as —.
Prices in every group
The same model costs different amounts in different groups; a key's group decides its line and its price.
| Group | Input ¥ per 1M tokens | Output ¥ per 1M tokens | Cache read ¥ per 1M tokens |
|---|---|---|---|
| Official · Enterprise | ¥87.6 | ¥438 | ¥2.19 |
| Claude Direct | ¥60 | ¥300 | ¥1.5 |
| Claude List Price | ¥96 | ¥480 | ¥2.4 |
| Claude PlusCheapest | ¥32 | ¥160 | ¥0.8 |
| Group | Input ¥ per 1M tokens | Output ¥ per 1M tokens | Cache read ¥ per 1M tokens |
|---|---|---|---|
| Standard | ¥24 | ¥48 | ¥0.802 |
| China ModelsCheapest | ¥16.8 | ¥33.6 | ¥0.561 |
How to choose
These points use each model's cheapest-group prices and catalog data. Before going live, run real requests with a key in your target group and check quality, speed and billing.
- On unit prices alone (each at its cheapest group): deepseek-v4-pro is cheaper on input, output and cache reads, so the same token usage costs less.
- When you resend the same long prefix (a fixed system prompt, the same long document), cached input is billed at the cache-read price: claude-fable-5-2 ¥0.8, deepseek-v4-pro ¥0.561 (per 1M tokens).
- Both support Chat Completions, so requests look the same and switching means changing the model parameter (the key's groups must offer both models).
- To call through Claude Messages: of the two, only claude-fable-5-2 supports it on this site.
- If production needs a known request source, use official or cloud groups. claude-fable-5-2: Official · Enterprise, Claude Direct and Claude List Price; deepseek-v4-pro: none on this site.
How to call them
Both models work with a Tutujin API key: set your SDK's base_url to https://www.tutujin.com/v1 and use the model name; everything else is written as with the official SDK.
- claude-fable-5-2: prices in every group, availability and API docs
- deepseek-v4-pro: prices in every group, availability and API docs
- Every Anthropic model and its price
- Every DeepSeek model and its price
- Chat Completions API reference
- Claude Messages API reference
- Quickstart: create a key and send your first request
- How to choose a group
- Models page: every model with prices in every group
- How billing works: pay as you go, groups and balance
More comparisons
FAQ
Which is cheaper, claude-fable-5-2 or deepseek-v4-pro?
In each model's cheapest group on this site, claude-fable-5-2 costs ¥32 input and ¥160 output (Claude Plus group) and deepseek-v4-pro costs ¥16.8 input and ¥33.6 output (China Models group), per 1M tokens. deepseek-v4-pro is cheaper on input, output and cache reads, by 48%, 79% and 30% respectively. What you actually pay depends on the split between input and output tokens and on the group you use. Prices updated Nov 13, 2026, 6:27 PM.
How do I call claude-fable-5-2 and deepseek-v4-pro?
Create an API key in the console, set your SDK's base_url to https://www.tutujin.com/v1 and use claude-fable-5-2 or deepseek-v4-pro as the model. claude-fable-5-2 supports: Chat Completions and Claude Messages. deepseek-v4-pro supports: Chat Completions.
Which groups offer claude-fable-5-2 and deepseek-v4-pro?
claude-fable-5-2: Official · Enterprise, Claude Direct, Claude List Price and Claude Plus. deepseek-v4-pro: Standard and China Models. No group offers both: give one key several groups (up to 5, in priority order) or use two keys.
Where do the prices on this page come from?
From our live price list — the same data as the model pages and the models page — shown for each model's cheapest group. Prices updated Nov 13, 2026, 6:27 PM.