Introducing G1! The New AI Frontier Model From G1 Max.G1
The frontier, from G1 Max. Designed for speed, built for reasoning.
Introducing G1! The New AI Frontier Model From G1 Max.Founders
Harvey & Mark
Benchmark Performance
G1:0.3b is a highly compact, 300M parameter model optimized specifically for zero-latency local execution and low-resource edge runtimes.
Context Window
Equipped with a 32K token context window. Seamlessly analyze long-context logs, complex source files, and maintain long-turn developer conversation state.
MoE Architecture
Leverages a sparse Mixture-of-Experts (MoE) architecture. Achieves sub-millisecond initial time-to-first-token (TTFT) and scales smoothly to multi-modal reasoning.