RecursiveSI/flint
main boardmodel
Flint is Recursive’s mixture-of-experts language model for coding agents, with 39.8 billion total parameters and approximately 2.5 billion active per token. Mixture-of-experts, 40B total (256 experts, 6 per token). Context: 256K tokens (config). License: apache-2.0.
- Org
- RecursiveSI
- Released
- 2026-10-21
- Status
- main board
- License
- apache-2.0 (permissive)
- Parameters
- 39.8B
- Architecture
- MoE · 256 experts, 6/token · FlintForCausalLM
- Context
- 262,144 tokens
- Task
- text-generation · text
- Signal
- 819 HF likes · 6444 downloads
Runs on
real quant files; need = weights + FP16 KV cache for 8K tokens + 1GB. A sizing calculation, not tested.
- Q4 ~26.9 GB: 32GB GPU (5090), Mac 36GB, Mac 48GB
- Q8 ~44.1 GB: 48GB GPU (RTX 6000 Ada / A6000), Mac 64GB (M1 Max), Mac 96GB, CPU only, 64GB RAM (slow)
- FP16 ~81.4 GB: 96GB GPU (RTX Pro 6000 Blackwell), 1× DGX Spark (128GB), 2× DGX Spark (256GB), Strix Halo 128GB unified (~96GB usable), Mac 128GB, Mac 192GB, Mac 256GB, Mac 512GB, CPU only, 128GB RAM (slow), CPU only, 256GB RAM (slow)
Links
- Hugging Face
- Ollama
- GGUF: bartowski/RecursiveSI_flint-GGUF
- GGUF: ggml-org/Flint-GGUF
- GGUF: lmstudio-community/flint-GGUF
- MLX: mlx-community/flint-OptiQ-4bit
- GGUF: unsloth/flint-GGUF
First published here 2026-10-21 · updated 2026-10-21 · markdown · JSON