GLM-5-3 (Flagship) vs NVIDIA Nemotron Labs Dory 1.0
A direct, empirical evaluation of GLM-5-3 (Flagship) (Zhipu AI & THUDM) against NVIDIA Nemotron Labs Dory 1.0 (NVIDIA).
GLM-5-3 (Flagship)
Developed by Zhipu AI & THUDM · 744B parameters
Complex algorithmic reasoning, multi-step code synthesis, and long-horizon tool execution.
NVIDIA Nemotron Labs Dory 1.0
Developed by NVIDIA · 4 Billion Parameters (Looped Transformer) parameters
Ultra-low VRAM memory residency (4.4 GB FP8), 256k long-context needle retrieval, and recurrent layer depth execution.
| Metric | GLM-5-3 (Flagship) | NVIDIA Nemotron Labs Dory 1.0 |
|---|---|---|
| SWE-bench Verified | 54.8% | 66.4% |
| MMLU-Pro | 78.4% | 82.8% |
| MATH-500 | 93.6% | 85.6% |
| Context Window | 256,000 tokens | 256,000 tokens |
| Input Pricing (per 1M) | $0.55 | $0.15 |
| Output Pricing (per 1M) | $2.19 | $0.45 |
Architectural & Practical Deployment Analysis
Evaluating GLM-5-3 (Flagship) (Zhipu AI & THUDM) alongside NVIDIA Nemotron Labs Dory 1.0 (NVIDIA) highlights key technical trade-offs across neural parameter scale and compute efficiency. GLM-5-3 (Flagship) executes across 744B total parameters with 44B active per token, whereas NVIDIA Nemotron Labs Dory 1.0 operates on 4 Billion Parameters (Looped Transformer) parameters with 4.1B Active Recurrent Block (Equivalent to 16B Depth) active per token.
On SWE-bench Verified software engineering tasks, GLM-5-3 (Flagship) achieves 54.8% autonomous resolution rate, compared to 66.4% for NVIDIA Nemotron Labs Dory 1.0. On complex reasoning benchmarks, GLM-5-3 (Flagship) posts 78.4% on MMLU-Pro and 93.6% on MATH-500, against 82.8% and 85.6% respectively for NVIDIA Nemotron Labs Dory 1.0.
From an operational economics perspective, processing 10,000,000 mixed tokens (7M input / 3M output) costs approximately $10.42 on GLM-5-3 (Flagship) versus $2.40 on NVIDIA Nemotron Labs Dory 1.0.
Recommendation: Deploy GLM-5-3 (Flagship) when your infrastructure prioritizes complex algorithmic reasoning, multi-step code synthesis, and long-horizon tool execution.. Deploy NVIDIA Nemotron Labs Dory 1.0 when workload requirements favor ultra-low vram memory residency (4.4 gb fp8), 256k long-context needle retrieval, and recurrent layer depth execution..