Controls creativity. 0 = predictable responses, higher = more varied and unexpected results. View Documentation
Type
FLOAT
Default
1.00
Range
0.00 ~ 2.00
Response Diversitytop_p
Filters responses to the most likely words within a probability range. Lower values keep answers more focused. View Documentation
Type
FLOAT
Default
1.00
Range
0.00 ~ 1.00
New Idea Encouragementpresence_penalty
Encourages or discourages using new words. Higher values promote fresh ideas, while lower values allow more reuse. View Documentation
Type
FLOAT
Default
0.00
Range
-2.00 ~ 2.00
Repetition Controlfrequency_penalty
Adjusts how often the model repeats words. Higher values mean fewer repeats, while lower values allow more repetition. View Documentation
Type
FLOAT
Default
0.00
Range
-2.00 ~ 2.00
Response Length Limitmax_tokens
Sets the max length of responses. Increase for longer replies, decrease for shorter ones. View Documentation
Type
INT
Default
--
Reasoning Depthreasoning_effort
Determines how much effort the model puts into reasoning. Higher settings generate more thoughtful responses but take longer. View Documentation
Type
STRING
Default
--
Range
low ~ high
Related Models
DeepSeek R1
deepseek-ai/DeepSeek-R1
DeepSeek-R1 is a reinforcement learning (RL) driven inference model that addresses issues of repetitiveness and readability within the model. Prior to RL, DeepSeek-R1 introduced cold start data to further optimize inference performance. It performs comparably to OpenAI-o1 in mathematical, coding, and reasoning tasks, and enhances overall effectiveness through meticulously designed training methods.
DeepSeek V3
deepseek-ai/DeepSeek-V3
DeepSeek-V3 is a mixture of experts (MoE) language model with 671 billion parameters, utilizing multi-head latent attention (MLA) and the DeepSeekMoE architecture, combined with a load balancing strategy that does not rely on auxiliary loss, optimizing inference and training efficiency. Pre-trained on 14.8 trillion high-quality tokens and fine-tuned with supervision and reinforcement learning, DeepSeek-V3 outperforms other open-source models and approaches leading closed-source models in performance.
DeepSeek R1 (Pro)
Pro/deepseek-ai/DeepSeek-R1
DeepSeek-R1 is a reinforcement learning (RL) driven inference model that addresses issues of repetitiveness and readability in models. Prior to RL, DeepSeek-R1 introduced cold start data to further optimize inference performance. It performs comparably to OpenAI-o1 in mathematical, coding, and reasoning tasks, and enhances overall effectiveness through carefully designed training methods.
DeepSeek V3 (Pro)
Pro/deepseek-ai/DeepSeek-V3
DeepSeek-V3 is a mixed expert (MoE) language model with 671 billion parameters, utilizing multi-head latent attention (MLA) and the DeepSeekMoE architecture, combined with a load balancing strategy without auxiliary loss to optimize inference and training efficiency. Pre-trained on 14.8 trillion high-quality tokens and fine-tuned with supervision and reinforcement learning, DeepSeek-V3 outperforms other open-source models and approaches leading closed-source models.
DeepSeek R1 Distill Llama 70B
deepseek-ai/DeepSeek-R1-Distill-Llama-70B
The DeepSeek-R1 distillation model optimizes inference performance through reinforcement learning and cold-start data, refreshing the benchmark for open-source models across multiple tasks.
