We use cookies to enhance your experience, provide personalised content, and analyse our traffic; see our Cookie Policy.
mistral/mistral-large-latestA European model at home working in French
Mistral Large is the flagship of Mistral AI, a French company. The mistral-large-latest alias currently points to Mistral Large 3, presented on 2 December 2025: a 675B-parameter mixture of experts with 41B active, under the Apache 2.0 licence, with 262,144 tokens of context.
According to Mistral it delivers best-in-class performance on multilingual conversations outside English and Chinese. Use it to write and edit in French, translate into and out of French, summarise French-language sources or drive agents with tool calls.
On Tchavi the product and support are in French. You pay with mobile money, no bank card and no KYC, with one API key for 80+ models. OpenAI-compatible: swap the base URL, keep your SDK. Your credits never expire.
Mistral AI, a French company; open weights, Apache 2.0
Multilingual performance beyond English and Chinese, per Mistral
Long French documents in a single call
Long texts in a single response
Drives agents and function-calling pipelines
Solve this step by step: a merchant sells 3 items at different prices with a discount on the total. Explain each calculation.
Summarise this article in 5 key points, then rewrite it as one newsletter paragraph, in simple, warm French.
Mistral Small (model: "mistral-small-latest") keeps the 262K of context and tools. Start with Small for everyday French text, and move up to Mistral Large when Small's quality is not enough. Test both on your own texts.
Devstral (model: "devstral-medium-latest") is Mistral's code-trained model, with 128K of context. To generate and refactor code, take Devstral; to write, translate and analyse in French, Mistral Large.
262,144 tokens of context in and up to 209,715 tokens out. Enough to handle a long French report in one go.
No, not on Tchavi: it takes text in, with no native JSON mode. Ask for the format in the prompt. For images or PDFs, pick Claude Sonnet 5 or GPT-6 Sol.
Swap the base URL for Tchavi's, use your Tchavi key and set model: "mistral-large-latest" on /v1/chat/completions. The alias follows the current Mistral Large; today, that is Large 3.
Yes. Top up your Tchavi account with Orange Money, MTN MoMo, M-Pesa or another mobile money operator. No bank card is needed. Credits land as soon as the payment is confirmed and are deducted as you use them.
Yes. Keep your current OpenAI SDK, just change the base URL to Tchavi's (shown in the API tab) and use your Tchavi key. The rest of your code stays the same.
import Tchavi from '@tchavi/sdk';
const client = new Tchavi({ apiKey: 'API_KEY' });
const response = await client.chat.completions.create({
model: 'mistral-large-latest',
messages: [
{
"role": "user",
"content": "Hello!"
}
],
});
console.log(response.choices[0].message.content);
console.log('Credits used:', response.tchavi.credits_used);Every parameter this model accepts on the API. Check the SDK for full typings.
modelRequiredID of the chat model (e.g. `mistral-large-latest`).
messagesRequiredConversation messages. Each item has a `role` (system/user/assistant) and a `content` that's either a plain string or an array of content blocks (`text`, `image_url`, `document_url`).
temperature1Sampling temperature between 0 and 2. Higher values = more random.
max_tokensMaximum number of tokens to generate in the response.
max_completion_tokensOpenAI alias for `max_tokens`. Either field works; if both are sent, `max_completion_tokens` wins.
streamfalseIf true, partial message deltas are streamed as Server-Sent Events.
toolsFunction-calling toolbox. Pass an array of `{ type: "function", function: { name, description, parameters } }`. Returned tool calls land in `choices[].message.tool_calls`. Only models with `supportsTools: true` accept this.
tool_choiceHow the model picks among tools: `"auto"` (default), `"none"` to force a plain text reply, or `{ type: "function", function: { name } }` to force a specific tool. Only with `supportsTools: true`.
top_p1Nucleus sampling: only tokens in the top `top_p` probability mass are considered.
seedInteger seed for deterministic output. The same seed + identical request yields the same completion when supported by the provider. OpenAI honours it; Anthropic ignores.
n1Number of completions to generate in one call. Defaults to 1; each extra completion is billed.
stream_optionsStreaming options, e.g. `{ "include_usage": true }` to receive a final SSE chunk with usage counts. Effective only when `stream: true`.
stopUp to 4 sequences where the API will stop generating further tokens.
frequency_penalty0Between -2.0 and 2.0. Positive values penalize tokens that have appeared frequently.
presence_penalty0Between -2.0 and 2.0. Positive values encourage the model to talk about new topics.
userEnd-user identifier passed to the upstream provider for abuse / moderation tracking. Should be stable per end-user.
logit_biasPer-token logit bias `{ "<tokenId>": <-100..100>, ... }`. OpenAI-specific; Anthropic ignores silently.
Estimated over 30 days, for a typical message: ~500 input + ~150 output tokens, billed to the thousandth of a credit.