Mistral Large Instruct 2411 is a frontier-class instruction model in the Mistral Large family, positioned by its publisher as strong on multilingual understanding, coding, and reasoning tasks. A third-party listing describes it as a 123-billion-parameter closed-source model released in the November 2024 wave of Mistral Large updates, with a Transformer-based architecture and a context window reported at 131K tokens. The Hugging Face identifier mistralai/Mistral-Large-Instruct-2411 is the canonical model reference, while IO.NET exposes the weights as a served endpoint so that teams can route requests through an inference provider rather than self-hosting. The framing as an instruct-tuned model signals that the system is shaped for direct conversational use, instruction following, and tool-augmented tasks rather than for raw base-model continuation.

Aggregated benchmark evidence gives a useful sense of how the model behaves in practice: on BFCL v4 the model posts a 38.4% score with a rank of 46, while on the Open LLM Leaderboard v2 it averages 46.5 and ranks 23 overall, with sub-scores of IFEval 84.01, BBH 52.74, MATH 49.55, GPQA 24.94, MUSR 17.22, and MMLU-Pro 50.69. The standout sub-score on IFEval points to reliable instruction following, while the balanced spread across reasoning, math, and knowledge benchmarks suggests a model that holds up across mixed workloads rather than excelling in a single niche. Practically, this makes Mistral Large Instruct 2411 a reasonable choice for production assistants and agent-style pipelines on IO.NET where a broad, generalist LLM with strong instruction-following behavior is more valuable than a narrow specialist, and where the larger context window supports multi-document summarization, code reviews, and longer conversational sessions.