Nemotron-Labs TwoTower 30B-A3B Base

Released
2026-06-25
Last refreshed
2026-07-02
Status
Researched 136d ago
Open weightsCommercial use: permittedCodingLong contextClassificationOpen Source

Nemotron-Labs TwoTower 30B-A3B Base is a released coding, long context, and classification model with open-weight and 128K context; evaluate it while provider pricing coverage matures.

NVIDIA AI releases · 27 in the last 12 months · this family litChangelog →
Specifications
Released
2026-06-25
Context
128K
Max output
128K
Parameters
~60B total checkpoint; Hugging Face reports 63B params
Architecture
MoE + SSM Hybrid
Knowledge cutoff
2025-06
Specialization
general
Openness
Open weights
License
NVIDIA Open ModelPermitted
Weights
Available
Code
Unknown
Training
Pretrained
Created by

Accelerated AI for enterprise solutions

Santa Clara, California, United States
Founded 2015
Website
Pricing

No tracked provider pricing is available yet.

About

Base text-generation checkpoint for Nemotron-Labs TwoTower. It uses a two-tower block-diffusion architecture over a Mamba-2/Transformer hybrid MoE backbone: a frozen causal AR/context tower processes clean prompt and committed tokens, while a trainable diffusion/denoiser tower fills token blocks by mask diffusion with cross-attention to the context tower. The checkpoint ships both towers and is not an instruction-tuned model.

Providers

No active provider routes are tracked for this model yet.

Capabilities

No model capability flags are currently sourced.

Benchmark peer barsfor Coding

Benchmark scores(10)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
78.25-shot, accuracy; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
60.95-shot, chain-of-thought exact match; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
92.7ARC-Challenge, 25-shot, acc_norm; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
76.15-shot, accuracy; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
88.90-shot, accuracy; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
75.60-shot; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
74.3MBPP-Sanitized, 3-shot; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
90.18-shot, accuracy; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
80.64-shot; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—
80.48-shot, average accuracy; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardObserved 2026-06-25—

Migration checks

No linked migration route is available for this model yet.

API versions

v1.1v1.0