Mistral
API key
9 params
Mistral Devstral Latest parameters
These are the parameters modelparams.dev tracks for Mistral Devstral Latest. Each row gives the type, default, valid range or values, and the conditions that gate it. It's the same data the JSON API serves.
| Parameter | Type | Default | Description | Condition |
|---|---|---|---|---|
| Length · 2 params | ||||
|
Max tokens
max_tokens
|
integer (1…+∞) | — | Maximum number of tokens to generate in the completion. | — |
|
Stop sequence
stop
|
string | — | Stops generation when this string is detected. | — |
| Sampling · 5 params | ||||
|
Temperature
temperature
|
number (0…1.5 step 0.1) | — | Controls randomness. Lower values make outputs more focused; higher values make them more varied. | — |
|
Top P
top_p
|
number (0…1 step 0.01) | 1 | Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. | — |
|
Random seed
random_seed
|
integer (0…+∞) | — | Seed used for deterministic sampling when reproducible outputs are desired. | — |
|
Presence penalty
presence_penalty
|
number (-2…2 step 0.1) | 0 | Penalizes repeated words or phrases to encourage a wider variety of generated content. | — |
|
Frequency penalty
frequency_penalty
|
number (-2…2 step 0.1) | 0 | Penalizes words based on how often they already appear in the generated text. | — |
| Output · 1 param | ||||
|
Response format
response_format.type
|
enum (text | json_object) | "text" | Controls whether the model returns normal text or JSON mode output. | — |
| Metadata · 1 param | ||||
|
Safe prompt
safe_prompt
|
boolean | false | Controls whether Mistral injects its safety prompt before the conversation. | — |