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 |
inclusionAI
|
|---|---|
| Company | inclusionAI |
| Release Date | 2026-08-20 |
| Parameters (B) | 124 |
| Architecture | Hybrid-linear MoE |
| Context Window | 262144 |
| Input Modalities | text |
| Open Source | Yes |
| License | mit |
| Score | 20 |
| VRAM (GB) | 256 |
| Compute | Full BF16 inference requires ~256GB VRAM (e.g., 4–8× H100/A100 80GB GPUs); FP8 quantization reduces requirements to ~128GB |
| Benchmark: AGIEval | 77.5 |
| Benchmark: ARC | 96.3 |
| Benchmark: BBH | 89.2 |
| Benchmark: C-Eval | 89.8 |
| Benchmark: HumanEval | 81.1 |
| Benchmark: LiveCodeBench | 40.1 |
| Benchmark: MATH-500 | 79 |
| Benchmark: MMLU-Pro | 67.7 |
| Benchmark: SimpleQA | 26.9 |
| Pricing (Input) | — |
| Pricing (Output) | — |