Models & Labs
NVIDIA NemotronBolt 74B-A9B runs at 2-bit
Sam Witteveenhigh confidence
Why it matters
- →Extreme 2-bit quantization is now viable for large reasoning models without catastrophic performance loss.
- →ParetoQ technique offers a new standard for balancing memory efficiency and model accuracy.
- →Enables high-concurrency local inference on consumer-grade hardware like RTX 5090.
NVIDIA has released NemotronBolt, a 74B-parameter model quantized to 2-bit precision using the ParetoQ technique. The model utilizes a specialized training pipeline involving pruning, post-training, and quantization-aware training with QAD (Quantization-Aware Distillation) to maintain performance. It is designed to run on consumer GPUs such as the RTX 5090, offering high concurrency and agentic tool-calling capabilities previously reserved for larger, more expensive hardware. The release includes weights on Hugging Face and a detailed research report outlining the methodology.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1NVIDIA Nemotron 3 Embed Tops RTEB Leaderboard
Hugging Face Blog · July 16, 2026 · Related
- 2NVIDIA Unveils Nemotron 3.5 Lightning and NeMo Switchyard
NVIDIA Blog · August 11, 2026 · Related
- 3NVIDIA Unveils Nemotron 3.5 Lightning
Lev Selector · August 14, 2026 · Background
- 4llama.cpp optimizes NVIDIA V100 inference performance
llama.cpp Releases · October 2, 2026 · Related
- 5llama.cpp b11372 optimizes Qwen4-Exp memory and GPU support
llama.cpp Releases · October 4, 2026 · Related
- 6Nemotron Fine-Tuning Hits Gold at IOI and IMO
Hugging Face Blog · October 7, 2026 · Related
- 7Nvidia releases open Nemotron models for agents and PCs
TechCrunch AI · October 19, 2026 · Related
- 8NVIDIA Dory: 4B Looped Transformer
Sam Witteveen · November 2, 2026 · Related
More in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
llama.cpp Releases
© TechCrunch AIModels & Labsmodels
Google deploys Gemini 4 Hydrogen for faster Search
Google is quietly shifting its search infrastructure by replacing older fast models with Gemini 4 Hydrogen, a tier designed specifically for speed and cost-efficiency rather than raw reasoning power. This model completes tasks 2.5 times faster than its predecessor while beating the more powerful Argon on specific intelligence benchmarks, proving that specialized efficiency can outperform brute force in high-volume workflows. By integrating this into Search's fan-out technique, Google now breaks down complex queries into subqueries with greater precision and speed, delivering richer visual responses without the latency of heavier models. This marks a strategic pivot toward agentic execution layers within consumer search, making AI Mode feel snappier for everyday multi-part questions.
TechCrunch AI