DeepSeek-V3.1-Terminus is an updated version of DeepSeek-V3.1 that addresses key user-reported issues while maintaining original capabilities. The model features improvements in language consistency and enhanced Code Agent and Search Agent performance.
Licensed under MIT, it offers unrestricted commercial use with the same model structure as DeepSeek-V3. Novita AI provides seamless cloud deployment for accessing this enhanced model through our scalable platform.
Access DeepSeek-V3.1-Terminus on Novita AI to leverage its improved capabilities without infrastructure complexity.
Current pricing on Novita AI: 131072 Context, $0.27/1M input tokens, $1/1M output tokens
Try DeepSeek V3.1 Terminus Demo
What is DeepSeek-V3.1-Terminus?
DeepSeek-V3.1-Terminus is an updated model that maintains the original capabilities of DeepSeek-V3.1 while addressing specific issues reported by users. Built on the DeepSeek-V3.1-Base model, it represents a refined version focused on practical improvements.
The model uses the same structure as DeepSeek-V3 and is released under the MIT License. This ensures open access for both research and commercial applications.
Key Improvements
Language Consistency
The update reduces instances of mixed Chinese-English text and eliminates occasional abnormal characters that appeared in previous versions.
Agent Capabilities
DeepSeek-V3.1-Terminus features optimized performance for:
- Code Agent: Enhanced coding capabilities and performance
- Search Agent: Updated template and tool-set with improved functionality
The Search Agent improvements are demonstrated in the provided assets/search_tool_trajectory.html file.
Performance Benchmarks
| Benchmark | DeepSeek-V3.1 | DeepSeek-V3.1-Terminus |
|---|---|---|
| Reasoning Mode w/o Tool Use | ||
| MMLU-Pro | 84.8 | 85.0 |
| GPQA-Diamond | 80.1 | 80.7 |
| Humanity’s Last Exam | 15.9 | 21.7 |
| LiveCodeBench | 74.8 | 74.9 |
| Codeforces | 2091 | 2046 |
| Aider-Polyglot | 76.3 | 76.1 |
| Agentic Tool Use | ||
| BrowseComp | 30.0 | 38.5 |
| BrowseComp-zh | 49.2 | 45.0 |
| SimpleQA | 93.4 | 96.8 |
| SWE Verified | 66.0 | 68.4 |
| SWE-bench Multilingual | 54.5 | 57.8 |
| Terminal-bench | 31.3 | 36.7 |
The benchmarks show significant improvements in key areas. Particularly noteworthy are Humanity’s Last Exam (36% improvement), BrowseComp (28% improvement), SimpleQA (3.6% improvement), and Terminal-bench (17% improvement).
Getting Started with DeepSeek-V3.1-Terminus on Novita AI
Accessing DeepSeek-V3.1-Terminus through Novita AI offers multiple pathways tailored to different technical expertise levels and use cases. Whether you’re a business user exploring AI capabilities or a developer building production applications, Novita AI provides the tools you need.
Use the Playground (Available Now – No Coding Required)
- Instant Access: Sign up and start experimenting with DeepSeek-V3.1-Terminus in seconds
- Interactive Interface: Test prompts and visualize outputs in real-time
- Model Comparison: Compare DeepSeek-V3.1-Terminus with other leading models for your specific use case
The playground enables you to test various prompts and see immediate results without any technical setup. It’s perfect for prototyping, testing ideas, and understanding model capabilities before full implementation.
Integrate via API (Live and Ready – For Developers)
Connect DeepSeek-V3.1-Terminus to your applications with Novita AI’s unified REST API.
Option 1: Direct API Integration (Python Example)
from openai import OpenAI
client = OpenAI(
base_url="https://api.novita.ai/openai",
api_key="your_api_key_here",
)
model = "deepseek/deepseek-v3.1-terminus"
stream = True # or False
max_tokens = 81920
system_content = "Be a helpful assistant"
temperature = 1
top_p = 1
min_p = 0
top_k = 50
presence_penalty = 0
frequency_penalty = 0
repetition_penalty = 1
response_format = {"type": "text"}
chat_completion_res = client.chat.completions.create(
model=model,
messages=[
{
"role": "system",
"content": system_content,
},
{
"role": "user",
"content": "Hi there!",
}
],
stream=stream,
max_tokens=max_tokens,
temperature=temperature,
top_p=top_p,
presence_penalty=presence_penalty,
frequency_penalty=frequency_penalty,
response_format=response_format,
extra_body={
"top_k": top_k,
"repetition_penalty": repetition_penalty,
"min_p": min_p
}
)
if stream:
for chunk in chat_completion_res:
print(chunk.choices[0].delta.content or "", end="")
else:
print(chat_completion_res.choices[0].message.content)Option 2: Multi-Agent Workflows with OpenAI Agents SDK
Build sophisticated multi-agent systems leveraging DeepSeek-V3.1-Terminus’s capabilities:
- Plug-and-Play Integration: Use DeepSeek-V3.1-Terminus in any OpenAI Agents workflow
- Advanced Agent Capabilities: Support for handoffs, routing, and tool integration
- Scalable Architecture: Design agents that leverage DeepSeek-V3.1-Terminus’s capabilities
Connect with Third-Party Platforms
Development Tools: Seamlessly integrate with popular IDEs and development environments like Cursor, Codex, Claude Code, Trae, Qwen Code and Cline through OpenAI-compatible APIs and Anthropic-compatible APIs.
Orchestration Frameworks: Connect with LangChain, Dify, CrewAI, Langflow, and other AI orchestration platforms using official connectors.
Hugging Face Integration: Novita AI serves as an official inference provider of Hugging Face, ensuring broad ecosystem compatibility.
Conclusion
DeepSeek-V3.1-Terminus delivers targeted improvements in language consistency and agent capabilities while maintaining the strong foundation of its predecessor. The enhanced Code Agent and Search Agent performance, combined with better language handling, makes it a practical choice for production applications.
Novita AI’s cloud platform provides easy access to these improvements without infrastructure overhead. Access DeepSeek-V3.1-Terminus on Novita AI today to leverage these enhancements in your AI applications.
Novita AI is an AI cloud platform that offers developers an easy way to deploy AI models using our simple API, while also providing the affordable and reliable GPU cloud for building and scaling.
