Model Detail

altar-1 is a large language model with 250.4B parameters released by AikidoSec. The model is registered under the text-generation pipeline tag on Hugging Face, distributed under a other license.

altar-1 ships with 250.4B parameters. Total weight footprint is approximately 500.8 GB, which is the relevant figure when planning local-inference VRAM. Distribution is governed by the other license — review the exact terms before commercial deployment.

altar-1 is best fit for general-purpose chat and instruction-following workloads. 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.