Liquid AI, released 2026-10-27. Dense hybrid decoder, 1.7B parameters, 256K context. Attention: ShortConv plus gated GQA with QK-Norm and sliding-window/global attention.
Dense hybridLiquid AI 🇺🇸 · May 28, 2026
LFM2.5 (1.2B)
Liquid AI's 1.2B LFM2.5 model keeps the 16-layer hybrid LIV convolution and GQA stack at a wider hidden size.
Key detail
Keeps the 16-layer LFM2.5 hybrid stack but doubles the hidden size to 2048 and uses 32 query heads.Parameters
1.2B
Context
128K
KV / token
12 KiB
Fact sheet
- Scale
- 1.2B parameters
- Context window
- 128,000 tokens
- Decoder type
- Dense hybrid
- Attention
- LIV convolution blocks plus GQA
- Layer mix
- 10 LIV conv + 6 GQA
- KV cache / token (bf16)
- 12 KiB · Very low
- License
- LFM Open License v1.0
- Released
Efficiency
Parameters active per token
100%
Dense: every weight runs for every token.
AA Intelligence Index v4.3.2
Source5.2
- Scientific
- 3.1
Architecture concepts
Sources & resources
Use the ← and → keys to step through releases.
Similar architectures
NVIDIA, released 2026-03-16. Dense hybrid decoder, 4B parameters, 256K context. Attention: GQA with only 4 attention layers.
Liquid AI, released 2026-05-28. Dense hybrid decoder, 350M parameters, 128K context. Attention: LIV convolution blocks plus GQA.
Bad Theory Labs, released 2026-07-20. Dense hybrid decoder, 27B parameters, 256K context. Attention: 3:1 Gated DeltaNet and Gated Attention.