# Together AI — Mamba-3

- Company: Together AI (together.ai)
- Announced: 2026-03-17T00:00:00+00:00
- Category: new-model
- Subject: Inference platform
- Models affected: Mamba-3, Mamba-2, Gated DeltaNet, Llama-3.2-1B
- Source: https://www.together.ai/blog/mamba-3
- Record: https://forck.live/items/2368-mamba-3

is a new state space model designed with inference efficiency as the primary goal, featuring a more expressive recurrence formula, complex-valued state tracking, and a MIMO variant that boosts accuracy without slowing down decoding. The model outperforms Mamba-2, Gated DeltaNet, and Llama-3.2-1B on prefill+decode latency at the 1.5B scale. The team open-sourced the kernels built using Triton, TileLang, and CuTe DSL.

## Evidence

Verbatim from https://www.together.ai/blog/mamba-3:

> Mamba-3 is a new state space model (SSM) designed with inference efficiency as the primary goal — a departure from Mamba-2, which optimized for training speed. The key upgrades are a more expressive recurrence formula, complex-valued state tracking, and a MIMO (multi-input, multi-output) variant that boosts accuracy without slowing down decoding. The result: Mamba-3 SISO beats Mamba-2, Gated DeltaNet, and even Llama-3.2-1B (Transformer) on prefill+decode latency across all sequence lengths at the 1.5B scale. The team also open-sourced the kernels, built using a mix of Triton, TileLang, and CuTe DSL for maximum hardware performance.

---

Record: https://forck.live/items/2368-mamba-3
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md