cinemasetfree
← Back to all generators
deepseek-ai/deepseek-v3.1
OfficialView on Replicate →
Latest hybrid thinking model from Deepseek
Capabilities
Max TokensTop-P
Cost
Community model (estimated from hardware time)
Input Parameters
| Name | Type | Description | Default | Constraints |
|---|---|---|---|---|
frequency_penalty | number | Frequency penalty | 0 | — |
max_tokens | integer | The maximum number of tokens the model should generate as output. | 1024 | — |
presence_penalty | number | Presence penalty | 0 | — |
prompt | string | Prompt | "Why are you better than Deepseek v3?" | — |
temperature | number | The value used to modulate the next token probabilities. | 0.1 | — |
thinking | string | Reasoning effort level for DeepSeek models. Use 'medium' for enhanced reasoning or leave as None for default behavior. | "None" | mediumNone |
top_p | number | Top-p (nucleus) sampling | 1 | — |
frequency_penaltynumberFrequency penalty
Default:
0max_tokensintegerThe maximum number of tokens the model should generate as output.
Default:
1024presence_penaltynumberPresence penalty
Default:
0promptstringPrompt
Default:
"Why are you better than Deepseek v3?"temperaturenumberThe value used to modulate the next token probabilities.
Default:
0.1thinkingstringReasoning effort level for DeepSeek models. Use 'medium' for enhanced reasoning or leave as None for default behavior.
Default:
"None"mediumNone
top_pnumberTop-p (nucleus) sampling
Default:
1Version:
133b8e778c13Updated: 8/13/2026485.6K runs