DeepSeek-V3.1-Terminus is an updated version of DeepSeek-V3.1 with enhanced language consistency, reduced mixed Chinese-English text, and optimized Code Agent and Search Agent performance.
On-demand deployments allow you to use DeepSeek V3.1 Terminus on dedicated GPUs with Fireworks' high-performance serving stack with high reliability and no rate limits.
DeepSeek V3.1 Terminus is an updated version of DeepSeek V3.1 developed by DeepSeek AI. This release improves language consistency (reducing mixed Chinese-English output), and enhances the performance of the Code Agent and Search Agent.
DeepSeek V3.1 Terminus is well-suited for:
The model supports a context length of 163,840 tokens on Fireworks.
The model has 685 billion parameters.
Yes. Fireworks supports fine-tuning this model using LoRA (Low-Rank Adaptation).
When deployed on-demand (dedicated GPUs), no rate limits apply.
DeepSeek V3.1 Terminus is released under the MIT License, allowing commercial use.
