Average Ratings 0 Ratings
Average Ratings 0 Ratings
Description
DeepSeek has launched DeepSeek-V3.1-Terminus, an upgrade to the V3.1 architecture that integrates user suggestions to enhance output stability, consistency, and overall agent performance. This new version significantly decreases the occurrences of mixed Chinese and English characters as well as unintended distortions, leading to a cleaner and more uniform language generation experience. Additionally, the update revamps both the code agent and search agent subsystems to deliver improved and more dependable performance across various benchmarks. DeepSeek-V3.1-Terminus is available as an open-source model, with its weights accessible on Hugging Face, making it easier for the community to leverage its capabilities. The structure of the model remains consistent with DeepSeek-V3, ensuring it is compatible with existing deployment strategies, and updated inference demonstrations are provided for users to explore. Notably, the model operates at a substantial scale of 685B parameters and supports multiple tensor formats, including FP8, BF16, and F32, providing adaptability in different environments. This flexibility allows developers to choose the most suitable format based on their specific needs and resource constraints.
Description
Runcrate offers AI teams and their agents a unified account to access GPUs and inference capabilities seamlessly. Users can rent various GPU models, including H100, H200, A100, L40S, RTX 4090, and RTX PRO 6000, with over 20 options available, for hourly use, featuring SSH, Jupyter, and VS Code directly in their browsers. A launch can be initiated in approximately four minutes, and users are only charged for running time, with most GPUs billing by the second. In terms of inference, clients can access more than 200 open models, such as DeepSeek, Qwen, Llama, Kimi, FLUX, and Whisper, through a single OpenAI-compatible API, ensuring compatibility with existing clients. Pricing is structured based on usage metrics like tokens, images, or seconds, while storage volumes allow for the retention of datasets and checkpoints across different instances. Reserved clusters can support over eight GPUs or provide month-long service options. Designed specifically for agents, users can connect tools like Claude Code, Codex, or Cursor to the Runcrate MCP server using a single command and sign in via their browser without needing to input any keys. Once connected, agents can easily launch a GPU, execute commands, and shut it down when finished. Additionally, the platform provides a command-line interface, which can be installed with a simple command, along with SDKs available for both Python and TypeScript to enhance development flexibility.
API Access
Has API
Yes
API Access
Has API
Yes
Screenshots View All
No images available
Integrations
ChatGPT
No
Claude
No
Claude Code
No
Claude Duet
No
Claude for Legal
No
Cursor
No
Hugging Face
Yes
Model Context Protocol (MCP)
No
OpenAI Codex
No
Visual Studio Code
No
Integrations
ChatGPT
Yes
Claude
Yes
Claude Code
Yes
Claude Duet
Yes
Claude for Legal
Yes
Cursor
Yes
Hugging Face
No
Model Context Protocol (MCP)
Yes
OpenAI Codex
Yes
Visual Studio Code
Yes
Pricing Details
Free
Free Trial
No
Free Version
Yes
Pricing Details
$0.65/hour
Free Trial
No
Free Version
No
Deployment
Web-Based
Yes
On-Premises
No
iPhone App
Yes
iPad App
Yes
Android App
Yes
Windows
No
Mac
No
Linux
No
Chromebook
No
Deployment
Web-Based
Yes
On-Premises
No
iPhone App
No
iPad App
No
Android App
No
Windows
Yes
Mac
Yes
Linux
Yes
Chromebook
No
Customer Support
Business Hours
No
Live Rep (24/7)
No
Online Support
Yes
Customer Support
Business Hours
No
Live Rep (24/7)
No
Online Support
Yes
Types of Training
Training Docs
Yes
Webinars
No
Live Training (Online)
No
In Person
No
Types of Training
Training Docs
Yes
Webinars
No
Live Training (Online)
No
In Person
No
Vendor Details
Company Name
DeepSeek
Founded
2023
Country
China
Website
api-docs.deepseek.com/news/news250922
Vendor Details
Company Name
Aeonmind LLC
Founded
2025
Website
www.runcrate.ai/