Token Plans
Plan team AI usage before buying credits
Token Plan is useful for AI coding and team workflows. API quota is better for product traffic. We separate those paths so customers understand what they are paying for.
Checked price references
Key Model Studio price references before a written quote
The full catalog now lives on the Models page. This pricing page keeps the buyer-oriented reference rows readable, while every model detail page shows its copied official detail-price summary where the console exposed one.
Catalog lanes
9
Flagship, Cost-optimized, Visual, Wan, Audio, Multimodal, Embeddings, Third-party, and Older.
Detail pages
90
84 detail-price summaries were cleanly matched; alias-style rows stay marked for console confirmation.
Full model catalog
Review every model detail
Open model catalogqwen3.7-max
Qwen3.7 flagship
Text-only flagship reasoning, coding, office, and autonomous agent work
$2.5 / 1M tokens
$7.5 / 1M tokens
International console original price. Limited-time discount may apply in the official console.
qwen3.7-plus
Qwen3.7 multimodal
Text, image, video input, GUI perception, coding, and tool-use workflows
$0.4 / 1M tokens
$1.6 / 1M tokens
International console original price for the <=256K input tier.
qwen3.6-plus
Qwen3.6 balanced
Vision-language production, coding, OCR-like extraction, and business automation
$0.5 / 1M tokens
$3 / 1M tokens
International console original price for the <=256K input tier.
qwen3.6-flash
Qwen3.6 cost-optimized
Fast multimodal production, support, extraction, code, and math workloads
$0.25 / 1M tokens
$1.5 / 1M tokens
International console original price for the <=256K input tier.
qwen3.6-max-preview
Qwen3.6 preview
Text-only preview lane for high-end coding and agent evaluation
$1.3 / 1M tokens
$7.8 / 1M tokens
International console original price for the <=128K input tier.
qwen3.6-27b
Qwen3.6 open source
Dense open-source vision-language lane for coding, STEM, and visual tasks
$0.6 / 1M tokens
$3.6 / 1M tokens
International console listed price.
deepseek-v4-pro
DeepSeek
Strong reasoning, code, math, research, and long-form technical analysis
$1.65 / 1M tokens
$3.301 / 1M tokens
Chinese Mainland console detail price captured for this model. International support and final quote must be confirmed.
deepseek-v4-flash
DeepSeek
Lightweight, high-concurrency reasoning and batch text processing
$0.138 / 1M tokens
$0.275 / 1M tokens
Chinese Mainland console detail price captured for this model. International support and final quote must be confirmed.
deepseek-v3.2
DeepSeek
Stable DeepSeek lane with sparse attention and reasoning/tool-use support
$0.287 / 1M tokens
Confirm output row
Chinese Mainland console detail price captured for input/cache rows. Confirm exact output tier in the official console before quoting.
kimi-k2.6
Kimi
Long-context code generation, instruction following, visual input, and agent dialogue
$0.8939 / 1M tokens
$3.7131 / 1M tokens
Chinese Mainland console detail price captured for this model.
glm-5.1
GLM
Long-horizon logic, summaries, code generation, and enterprise assistants
$0.825 / 1M tokens
$3.301 / 1M tokens
Chinese Mainland console detail price captured for the <=32K input tier. Other tiers, regions, and final quotes may differ.
MiniMax-M2.5
MiniMax
Agent-style coding, tool invocation, search, productivity, and office work
$0.304 / 1M tokens
$1.213 / 1M tokens
Chinese Mainland console detail price captured for this model.
Image generation / editing
qwen-image-2.0
$0.035 / image
Input
Text / image
Output
Image
High-quality image generation / editing
qwen-image-2.0-pro
$0.075 / image
Input
Text / image
Output
Image
Wan image generation / editing
wan2.7-image
$0.03 / image
Input
Text / image
Output
Image
Wan image generation / editing
wan2.7-image-pro
$0.075 / image
Input
Text / image
Output
Image
Text to video
happyhorse-1.0-t2v
$0.14/s 720P · $0.24/s 1080P
Input
Text
Output
Video
Image to video
happyhorse-1.0-i2v
$0.14/s 720P · $0.24/s 1080P
Input
Text / image
Output
Video
Reference to video
happyhorse-1.0-r2v
$0.14/s 720P · $0.24/s 1080P
Input
Text / image
Output
Video
Video editing
happyhorse-1.0-video-edit
$0.14/s 720P · $0.24/s 1080P
Input
Image / video
Output
Video
Embedding
text-embedding-v4
$0.07 / 1M tokens
Input
Text
Output
Vector
Reranking
qwen3-rerank
$0.1 / 1M tokens
Input
Text
Output
Ranked text
This section is intentionally a reference index, not a complete quote table. Token, cache, media, audio, video-second, and tool-call rows use different billing units, so final procurement review should start from the exact model detail page and end with a written quote.
Standard
USD 30
25,000 Credits
Light daily AI usage
Advanced
USD 100
100,000 Credits
Frequent AI coding and content work
Premium
USD 200
250,000 Credits
Core users who rely on AI throughout the day
Shared quota pack
USD 700
625,000 Credits
Elastic overage pool for teams
Token Plan pricing is separate from model invocation pricing above. Final quotes must verify official plan availability, billing currency, taxes, exchange rate, payment method, account eligibility, and service fees.
Billing structure
Separate official usage costs from procurement service fees
This is the most important trust point. Buyers should see the platform cost, service fee, payment cost, and delivery scope before payment.
New-user free quota
Check whether the buyer has unused trial allowance before purchasing credits.
Model invocation pricing
Confirm input and output token rates for the selected model and deployment region.
Training and deployment
Separate fine-tuning, deployment, and hosting costs from normal API usage.
Savings plans
Estimate whether volume commitments or prepaid packages make sense.
Bills and cost management
Review monthly usage, alerts, payment route, and replenishment timing.
Service scope
What ModelSmarter charges for
The customer should not confuse our service fee with official model usage cost. This section makes the service scope explicit.
Service
Model Review
$149
Service
Procurement Setup
From $790
Service
Managed Team
Custom
Quote process
Four checks before payment
A written quote should be the conversion point, not a generic checkout button.
01
Choose the model lane
Match workload, modality, latency needs, region, and context length before quoting.
02
Confirm billing route
Separate official model usage, Token Plan seats, shared quota, payment fees, and service scope.
03
Set up access
Coordinate account route, API key, base URL, endpoint region, and team handoff.
04
Monitor usage
Track consumption, rate limits, free quota, cost controls, and monthly replenishment needs.