DeepSeek: DeepSeek V3.1 Terminus

DeepSeek-V3.1 Terminus is an update to DeepSeek V3.1, maintaining its original capabilities while enhancing language consistency and agent performance, particularly in coding and search agents. It's a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes. It extends the DeepSeek-V3 base with a two-phase long-context training process, reaching up to 128K tokens, and uses FP8 microscaling for efficient inference. Users can control reasoning behavior with the `reasoning` `enabled` boolean. This model significantly improves tool use, code generation, and reasoning efficiency, achieving performance comparable to DeepSeek-R1 on difficult benchmarks while responding more quickly. It supports structured tool calling, code agents, and search agents, making it ideal for research, coding, and agentic workflows. It offers a 163K token context window and 4K token max output. Pricing is $0.21/0.79 per 1M tokens (input/output).