Last verified
ROLE PLAY128K CONTEXTTEXTCONTEXT CACHINGCNY SOURCE

Doubao Seed Character API Pricing

Doubao Seed Character is ByteDance's specialised character-interaction and multi-turn drama model. The official Volcano Ark table starts the <=32K tier at Y0.80/M input, Y2.00/M output, and Y0.16/M cache-hit input; AI//COST stores that as $0.119/M, $0.2976/M, and $0.0238/M at Y6.7209/USD. Pulled directly from volcengine.com daily.

Input - per 1M tokens
$0.119/M
Original Y0.80/M CNY source
Output - per 1M tokens
$0.298/M
Original Y2.00/M CNY source
Cached input - per 1M tokens
$0.0238/M
Original Y0.16/M -80%
Effective - agentic blend
$0.0615/M
92/8 split - 82% cache
§ 01 / TERMINAL

Run the numbers.

Live calculator pre-loaded with Doubao Seed Character's current public rate. Use it for character-chat planning, synthetic persona testing, and long multi-turn role-play budgeting.

$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
Open full calculator (all models · share URL · CSV) →
§ 02 / SCENARIOS

Real-world presets.

NPC CHAT

Game NPC dialogue

$0.0007/turn
4k in - 1.8k out~142,975 turns/$100
SIMULATION

Persona simulation

$0.00141/session
9k in - 3.5k out~70,921 sessions/$100
STORY

Interactive scene

$0.00099/scene
6k in - 2.5k out~101,049 scenes/$100
SUPPORT

Brand voice assistant

$0.00048/turn
3k in - 1.2k out~208,365 turns/$100
§ 03 / TOKENIZER

Paste text. See tokens. See cost.

Calibrated · measured on the vendor's tokenizer · 2026-06-10 Auto-counts as you type

Counts use a chars-per-token calibration measured on the vendor's own published tokenizer (ByteDance-Seed/Seed-OSS-36B-Instruct, 2026-06-10). English prose is typically within a few percent; code and non-Latin scripts tokenize heavier. For billing-exact counts use the vendor's count-tokens API.

Characters 605
Words 91
Tokens (estimated) 116 tokens
Cost as input · uncached $0.00001 USD
Cost as output · uncached $0.00003 USD
Cost as cached input $0.000003 USD
§ 04 / SHELF

Up against the shelf.

All models →
Model Input /M Output /M Effective blended Context Best for
Doubao Seed Character Current $0.119 cache $0.0238 $0.298 $0.0615 agentic 92/8 128K Persona and role-play dialogue
Doubao Seed 1.6 Flash $0.0223 cache $0.0045 $0.223 $0.0249 cheap alternative 256K High-volume low-cost chat
Doubao Seed 1.6 $0.119 cache $0.0238 $1.19 $0.133 general-purpose sibling 256K General thinking and multimodal chat
Doubao Seed 2.0 Mini $0.0298 cache $0.006 $0.298 $0.0333 active 2.0 budget peer 256K Fast low-latency routing
Kimi K2.5 $0.60 cache $0.10 $3.00 $0.415 cross-vendor creative peer 262K Kimi multimodal agents

Frequently asked.

Practical character-model pricing questions, with the public CNY ladder separated from workload assumptions.

Q · 01 What is Doubao Seed Character priced at? +
The <=32K baseline is Y0.80/M input, Y2.00/M output, and Y0.16/M cache-hit input. AI//COST stores that as $0.119/M, $0.2976/M, and $0.0238/M using Y6.7209/USD.
Q · 02 Does the price change above 32K input length? +
Yes. The 32K-128K tier moves to Y1.20/M input and Y6.00/M output, while cache-hit input stays at Y0.16/M on the public table. This page uses the baseline <=32K row for clean public comparisons.
Q · 03 What is this model best for? +
ByteDance positions it for character interactions and multi-turn drama. In practice that means NPC dialogue, branded persona simulations, and longer role-play sessions where voice consistency matters more than raw reasoning breadth.
Q · 04 How does it compare with generic Seed 1.6? +
Doubao Seed 1.6 uses the same baseline input price but a much higher output rate at $1.1903/M. Seed Character is cheaper when your workload is heavy on generated dialogue but does not need the broader general-purpose positioning.
Q · 05 Are taxes or private discounts included here? +
No. This page stores the public Volcano Ark list price before tax, enterprise rebates, or negotiated terms. The USD layer exists only to make cross-vendor comparisons easier.
Q · 06 Why does the effective figure look so low? +
Because AI//COST applies a standard 92/8 workload with an 82% cache-hit assumption when showing the blended planning figure. The actual vendor list price remains the CNY row shown above.
Y
Reviewed by Yaroslav Vikhariev Founder - AI//COST - Pricing pulled daily from volcengine.com - Last verified July 11, 2026
Methodology Report a correction More by Y.V.