LFM2.5-2.6B

Released2026-08-04Size2.7B denseArchitecturehybrid linear attnLicenselfm1.0AA rank#70 of 75

Files and memory at 8K context

Quant	File	Files	+ KV + runtime	Smallest fit	Source
Q4_K_M LiquidAI/LFM2.5-2.6B-GGUF	1.56 GiB	1	2.7 GiB+	8 GB GPU	HF files
Q5_K_M LiquidAI/LFM2.5-2.6B-GGUF	1.81 GiB	1	2.9 GiB+	8 GB GPU	HF files
Q8_0 LiquidAI/LFM2.5-2.6B-GGUF	2.68 GiB	1	3.8 GiB+	8 GB GPU	HF files
FP16 LiquidAI/LFM2.5-2.6B-GGUF	5.03 GiB	1	6.2 GiB+	8 GB GPU	HF files
MLX 4-bit mixed precision LiquidAI/LFM2.5-2.6B-MLX-4bit	1.47 GiB	1	2.6 GiB+	8 GB Mac	HF files
MLX 8-bit LiquidAI/LFM2.5-2.6B-MLX-8bit	2.67 GiB	1	3.8 GiB+	8 GB Mac	HF files

Sizes are exact file bytes (text weights; projectors listed separately). “+” = recurrent state or other terms unknown. Smallest fit adds 0.5 GiB spare; Macs use 75% of memory.

Quality

Artificial Analysis Intelligence Index

8.4 AA estimate

Variant: LFM2.5-2.6B · observed 2026-10-17 · source

Rank #70 of 75 researched 2026 models.

Arena text rating

Unknown

Scores describe the evaluated hosted variant, not each quantized file. AA and Arena are never averaged.

Reported speeds 0

No reported measurement yet. Submit one.

Community threads

<|tool_call_start|> and <|tool_call_end|> The biggest problem with this model. Hugging Face · 2026-08-27
Internal model structure preview Hugging Face · 2026-09-06

Qualitative signals only; titles are not verified claims.