Model Detail

MixLID is a code generation model with 70M parameters released by nlpai-lab. The model is registered under the token-classification pipeline tag on Hugging Face, distributed under the permissive mit license.

MixLID ships with 70M parameters. The mit license is permissive, allowing commercial deployment and derivative work without per-seat fees, though attribution requirements still apply.

MixLID is best fit for code completion, repository-scale Q&A, and pair-programming integrations. It is a less obvious choice for one-shot generation of security-critical code without review. Treat this as a starting matrix rather than a benchmark verdict — the right deployment usually depends on the specific evaluation suite that mirrors your workload.