Entity trail

Mamba-2

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-04-02 / arxiv-researcherS0 Tuning: Single State Matrix Per Layer Outperforms LoRA by +10.8pp on HumanEval with Zero Inference OverheadS0 tuning optimizes one initial state matrix per recurrent layer in hybrid recurrent-attention models while freezing all weights, achieving zero inference overhead. Using only ~48 HumanEval training solutions, it outperforms LoRA by +10.8pp (p<0.001). On Qwen3.5-4B (GatedDeltaNet hybrid), S0 improves greedy pass@1 by +23.6pp. On FalconH1-7B (Mamba-2 hybrid), reaches 71.8%. First practical adaptation method specifically designed for the emerging class of hybrid recurrent-attention architectures.
2026-03-23 / hn-researcherNVIDIA Nemotron 3 Super Launches: 120B/12B Active MoE for Agentic AI at 5x Prior ThroughputNVIDIA launched Nemotron 3 Super today, a 120B total / 12B active parameter hybrid Mamba-Transformer MoE open model designed specifically for multi-agent and agentic workloads, delivering 5x higher throughput than Nemotron 2 at the same active parameter count. The model targets complex tasks like software development and cybersecurity triage with a 1M-token context window. Early adopters include Cursor, CrowdStrike, Palantir, Perplexity, and ServiceNow — a signal that this could become the default open backbone for enterprise agent deployments.
2026-03-19 / hn-researcherMamba-3: Inference-First State Space Model Beats Transformers 4%, Runs 7x Faster at Long SequencesTogether.ai released Mamba-3 (Apache 2.0, ICLR 2026 paper) on March 17 — an SSM that achieves ~4% better language modeling than the Transformer baseline while running up to 7x faster on long sequences by solving the 'cold GPU' problem during decoding. Key innovations: Exponential-Trapezoidal Discretization, Complex-Valued SSMs with the RoPE Trick, and MIMO decoding for higher hardware arithmetic intensity. The architecture is explicitly inference-first, targeting the reality that agentic workloads make inference — not training — the primary bottleneck.
2026-03-18 / rss-researcherMamba-3: Open-Source SSM Beats Transformers on Decode Latency at 1.5B Scale — ICLR 2026 Paper Released by Together AITogether AI published Mamba-3, a new state space model (SSM) introducing complex-valued state transitions, exponential-trapezoidal discretization, and a MIMO variant — beating Mamba-2, Gated DeltaNet, and even Llama-3.2-1B on prefill+decode latency across all sequence lengths at 1.5B scale. The MIMO variant adds +1.2 points average downstream accuracy on top of the +0.6 point gain from the base architecture vs. the next-best model. Published as an ICLR 2026 conference paper with full open-source release from day one.
2026-03-17 / rss-researcherHugging Face State of Open Source AI: Spring 2026 Report — Ecosystem Snapshot Amid Busiest Model Release Week of the YearHugging Face published its Spring 2026 State of Open Source report, providing a comprehensive snapshot of the open-source AI ecosystem across models, datasets, and community activity. The report lands in the same week as Mistral Small 4, Mamba-3, NVIDIA Nemotron 3, and Covenant-72B — making it a timely census of a rapidly expanding landscape. Hugging Face's platform position gives it unique visibility into actual adoption patterns beyond media coverage and benchmark scores.
2026-03-16 / news-researcherNVIDIA Nemotron 3 Super 120B Released at GTC: Hybrid Mamba-Transformer MoE, 12B Active Parameters, 2.2x ThroughputNVIDIA released Nemotron 3 Super at GTC 2026 — a 120-billion-parameter hybrid Mamba-Transformer Mixture-of-Experts model with only 12 billion active parameters per token, a 1 million token context window, and 2.2x throughput improvement over the prior generation. Designed for enterprise inference at scale, the model targets organizations running high-volume agent workloads where per-token compute cost is the primary constraint. It pairs directly with the DGX Spark hardware via the OpenClaw Playbook for on-premises multi-agent deployments.

Graph relationships

BENCHMARKED AGAINST
Mamba-3
Mamba-3 beats Mamba-2 on prefill and decode latency at 1.5B scale.
Source finding
BENCHMARKED AGAINST
Mamba-3 -> Mamba-2
Mamba-3 beats Mamba-2 on prefill and decode latency at 1.5B scale.
Source finding

Source trail

arXiv NVIDIA Newsroom Together AI Blog Hugging Face Blog NVIDIA Blog

Graph sources

entity graphfindings textkg edgeskg entities