Compare LLM Models
Select 2-5 models to compare side by side.
Select models to compare
0 / 5 selected
Top downloads
Select 2-5 models
| Attribute |
LGAI-EXAONE
|
|---|---|
| Company | LGAI-EXAONE |
| Release Date | 2026-07-31 |
| Parameters (B) | 750 |
| Architecture | MoE Transformer |
| Context Window | 262144 |
| Input Modalities | text |
| Open Source | Yes |
| License | apache-2.0 |
| Score | 20 |
| VRAM (GB) | 750 |
| Compute | Recommended deployment on a single node with 8× NVIDIA H200 GPUs (tensor parallelism 8); FP8 weights ~750 GB, with SGLang or vLLM and FP8 KV cache |
| Benchmark: AIME-2026 | 92.3 |
| Benchmark: GPQA-Diamond | 82.2 |
| Benchmark: HLE | 18.3 |
| Benchmark: IFEval | 92.4 |
| Benchmark: MMLU-Pro | 83.5 |
| Benchmark: SWE-Bench-Verified | 68.2 |
| Benchmark: Terminal-Bench-2.1 | 43.8 |
| Pricing (Input) | — |
| Pricing (Output) | — |