SenseChat-Turbo

SenseChat-Turbo

SenseTime’s “Ri Ri Xin” Series of Lightweight Language Models
2024-04-23
LLM
Input:
$0.33/1M tokens
Output:
$0.77/1M tokens
Bulk order? Contact your manager for exclusive deals

API Overview

SenseChat-Turbo (Shangshang Turbo) is a high-performance, cost-effective version within SenseTime’s “Daily New SenseNova” large model ecosystem. Its core positioning is **“ultra-high inference efficiency + industrial-grade task alignment,”** aiming to provide enterprises and developers with a productivity tool that strikes the perfect balance between performance and cost, primarily targeting GPT-4 Turbo.

  • Outstanding Inference Performance: Built on the Daily New 5.0/5.5 system and optimized through a Mixture-of-Experts (MoE) architecture, it significantly boosts generation speed and reduces latency, enabling real-time streaming text output.
  • Ultra-long Context Window of 200K Tokens: It supports context windows of up to 200K tokens (approximately 150,000 Chinese characters), allowing it to process entire books, lengthy legal contracts, or massive codebases in one go, making it exceptionally powerful for long-text analysis and information extraction.
  • Deep Instruction Following: Specifically enhanced for complex business logic and multi-level instructions, it can precisely execute structured outputs (such as JSON format), write long-form texts, and perform rigorous logical reasoning.
  • Cost-Effective and Scalable: Compared to the Reasoner series, the Turbo version dramatically reduces token usage costs, making it ideal for high-frequency, concurrent commercial applications such as intelligent customer service, content creation pipelines, and large-scale knowledge-base question-answering systems.
  • Full-Stack Domestic Adaptation: Deeply optimized for domestically produced computing infrastructures, ensuring stable operation on SenseCore SenseTime AI platforms, and supporting private deployment as well as rapid API integration.

───────────────────────────────────────────────────────────────────

Core Capabilities

Ultra-Fast Response: Features extremely high first-token output speed, making it suitable for highly real-time interactive dialogue applications and real-time analysis scenarios.

📝 Creative and Copywriting Master: Outstanding in humanities-related tasks such as creative writing, news稿 generation, and email drafting, with a more natural language style that closely matches human expression habits.

💻 Code Assistant: Equipped with powerful code understanding and generation capabilities, it supports logical coding, bug fixing, and unit test generation for mainstream programming languages, boosting development efficiency.

🗂️ Massive Information Processing: Can quickly retrieve keywords, summarize key ideas, and establish logical connections across documents from vast amounts of unstructured data.

📊 Business Process Automation: Through Function Calling capabilities, it can act as a “brain” to connect with external tools, automating tasks such as ticket booking, data queries, and business approval workflows.

Playground

Log in to explore more features! Click to Log In

API Analytics

API Reference (1)

API DescriptionAPI EndpointRequest MethodStabilityParameter Description
Chat (SenseTime)
POST
Stable
View Details

API Pricing

$
ModelDescriptionContextOfficial Price302.AI Price

SenseChat-Turbo

-
32000

Input$0.3 / 1M tokens
Output$0.7 / 1M tokens

Input$0.33/ 1M tokens
Output$0.77/ 1M tokens
10%