Flint AWQ is an open-source language model by mattbucci. Features: 39.8b LLM, VRAM: 0.9GB, Context: 256K, License: apache-2.0, Quantized, LLM Explorer Score: 0.3.
| LLM Name | Flint AWQ |
| Repository 🤗 | https://huggingface.co/mattbucci/Flint-AWQ |
| Base Model(s) | |
| Model Size | 39.8b |
| Required VRAM | 0.9 GB |
| Updated | 2026-10-26 |
| Maintainer | mattbucci |
| Model Type | flint |
| Model Files | |
| AWQ Quantization | Yes |
| Quantization Type | awq |
| Model Architecture | FlintForCausalLM |
| License | apache-2.0 |
| Context Length | 262144 |
| Model Max Length | 262144 |
| Transformers Version | 5.12.1 |
| Tokenizer Class | PreTrainedTokenizerFast |
| Vocabulary Size | 150144 |
| Torch Data Type | bfloat16 |
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Flint Mixed 3 4bit | 256K / 17.9 GB | 1609 | 3 |
| Flint MLX Serve 4bit | 256K / 22.5 GB | 1950 | 2 |
| Flint MLX Serve Mixed 4 8bit | 256K / 23.4 GB | 1230 | 3 |
| Flint Mixed 3 4bit | 256K / 17.9 GB | 193 | 1 |
| Flint | 256K / 80 GB | 4795 | 754 |
| Flint FP8 BF16 | 256K / 0.8 GB | 46 | 1 |