LFM
32 models34 ranked conversions252K downloads
Best conversion per model
One row per base model. The full list of attempts is on each model page.
| # | Model | Author | Params | KL | Refusals | Score | Evidence | Downloads |
|---|---|---|---|---|---|---|---|---|
| 1 | LFM2.5-1.2B-Thinking LiquidAI | 0xzknw | 1.2B | 0.0003 | 3/1003% | 98.5 | BestParetoGGUFSAFETENSORS | 614 |
| 2 | LFM2.5-8B-A1B LiquidAI | zaakirio | 8B · 1B active | 0.048 | 0/1000% | 97.7 | BestParetoGGUFSAFETENSORS | 8.5K |
| 3 | LFM2-24B-A2B LiquidAI | trohrbaugh | 24B · 2B active | 0.055 | 0/1000% | 97.4 | BestParetoFP8GGUFSAFETENSORS | 6.5K |
| 4 | LFM2-VL-3B LiquidAI | ZuzeTt | 3B | 0.020 | 4/1004% | 97.0 | BestParetoGGUFSAFETENSORS | 14.9K |
| 5 | LFM2.5-1.2B-Instruct LiquidAI | saidutta69 | 1.2B | 0.066 | 2/1002% | 95.9 | BestParetoGGUFSAFETENSORS | 1.2K |
| 6 | LFM2.5-2.6B LiquidAI | atakankutalp | 2.6B | 0.023 | 6/1006% | 95.9 | BestParetoGGUFLORAMLX | 2.2K |
| 7 | LFM2.5-VL-1.6B LiquidAI | coder3101 | 1.6B | 0.011 | 8/1008% | 95.4 | BestParetoGGUFMLXSAFETENSORS | 4K |
| 8 | LFM2-350M LiquidAI | MihaiPopa-1 | 0.4B | 0.042 | 6/1006% | 95.0 | BestParetoSAFETENSORS | 114 |
| 9 | LFM2.5-350M LiquidAI | coder3101 | 0.4B | 0.044 | 6/1006% | 94.9 | BestParetoGGUFMLXSAFETENSORS | 2.5K |
| 10 | LFM2.5-VL-450M LiquidAI | coder3101 | 0.4B | 0.017 | 9/1009% | 94.7 | BestParetoGGUFSAFETENSORS | 3.8K |
| 11 | LFM2-1.2B-RAG LiquidAI | cochbild | 1.2B | 0.090 | 6/1006% | 92.9 | BestParetoGGUFSAFETENSORS | 4.6K |
| 12 | LFM2.5-230M LiquidAI | saidutta69 | 0.2B | 0.055 | 10/10010% | 92.4 | BestParetoGGUFSAFETENSORS | 724 |
| 13 | LFM2-700M LiquidAI | spitfire4794 | 0.7B | 0.123 | 8/1008% | 90.5 | BestParetoGGUFSAFETENSORS | 4.6K |
| 14 | lfm2.5-2.6b-fable5-coding-agent AyoubChLin | saidutta69 | 2.6B | 0.014 | 96/10096% | 51.3 | BestParetoGGUFSAFETENSORS | 29.9K |
| 15 | LFM2.5-24B-A2B LiquidAI | OS-Software | 24B · 2B active | 0.017 | 99/10099% | 49.6 | BestParetoGGUFGPTQSAFETENSORS | 590 |
All models
LFM2.5-1.2B-Thinking1.2BLFM2.5-8B-A1B8B · 1B activeLFM2-24B-A2B24B · 2B activeLFM2-VL-3B3BLFM2.5-1.2B-Instruct1.2BLFM2.5-2.6B2.6BLFM2.5-VL-1.6B1.6BLFM2-350M0.4BLFM2.5-350M0.4BLFM2.5-VL-450M0.4BLFM2-1.2B-RAG1.2BLFM2.5-230M0.2BLFM2-700M0.7Blfm2.5-2.6b-fable5-coding-agent2.6BLFM2.5-24B-A2B24B · 2B activeHuihui-LFM2-2.6B-Exp-abliterated2.6BLFM2-2.6B-Exp-heretic2.6BLFM2-24B-A2B-heretic-ayun23.8BLFM2-8B-A1B8B · 1B activeLFM2-8B-A1B-Abliterated-MLX-4bit8B · 1B activeLFM2-VL-1.6B1.6BLFM2.5-1.2B-Base1.2BLFM2.5-1.2B-Instruct-heretic1.2BLFM2.5-1.2B-Thinking-Gemini-Pro-Heretic-Uncensored-DISTILL1.2BLFM2.5-1.2B-Thinking-GPT-5.1-1000-Heretic-Uncensored-DISTILL1.2BLFM2.5-1.2B-Thinking-Heretic-NX-Prime-v21.2BLFM2.5-1.2B-Thinking-Polaris-Heretic-Uncensored-DISTILL1.2BLFM2.5-2.6B-heretic2.6BLFM2.5-230M-GGUF—LFM2.5-24B-A2B24B · 2B activeLFM2.5-8B-A1B-Base8B · 1B activeLFM2.5-VL-3B3B
Unmeasured models
Huihui-LFM2-2.6B-Exp-abliterated, LFM2-2.6B-Exp-heretic, LFM2-24B-A2B-heretic-ayun, LFM2-8B-A1B, LFM2-8B-A1B-Abliterated-MLX-4bit, LFM2-VL-1.6B, LFM2.5-1.2B-Base, LFM2.5-1.2B-Instruct-heretic, LFM2.5-1.2B-Thinking-Gemini-Pro-Heretic-Uncensored-DISTILL, LFM2.5-1.2B-Thinking-GPT-5.1-1000-Heretic-Uncensored-DISTILL, LFM2.5-1.2B-Thinking-Heretic-NX-Prime-v2, LFM2.5-1.2B-Thinking-Polaris-Heretic-Uncensored-DISTILL, LFM2.5-2.6B-heretic, LFM2.5-230M-GGUF, LFM2.5-24B-A2B, LFM2.5-8B-A1B-Base, LFM2.5-VL-3B — conversions exist but their publishers did not report KL and refusal numbers, so they are not ranked.