LFM2.5 2.6B — Hardware Requirements & GPU Compatibility
ChatFunctions
LFM2.5-2.6B is Liquid AI's on-device model in the LFM2.5 family, a hybrid architecture mixing gated short-convolution blocks with grouped-query attention layers, at about 2.7 billion parameters. It is post-trained for agentic workloads, tool use, data extraction, retrieval-augmented generation, and long-context tasks, though the publisher does not recommend it for heavy agentic coding or knowledge-intensive question answering. It supports a 131K token context window and is released under Liquid AI's custom LFM1.0 license. At around 2.7 billion parameters, it runs easily on almost any modern laptop or consumer GPU at 4-bit quantization, and the publisher reports it using well under 2.5 GB of memory during on-device inference.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 128K)Best on 24 GB (Q4_K_M · ~321 tok/s · 128K)Macs from 8 GB (Q4_K_M · ~35 tok/s)
Other
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run LFM2.5 2.6B
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
ollama run hf.co/LiquidAI/LFM2.5-2.6B-GGUF:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 21.7 GB free · 16K context
Download: LFM2.5-2.6B-Q4_K_M.gguf1.67 GBLiquidAI
Other ways
each opens step-by-stepRun in cloud
Fits on RTX 3060 12GB (9 GB headroom) · Q4_K_M
- Generation speed
- ~115 tok/s
- generation speed
- Cost per 1M output tokens
- $0.20
- per 1M output tokens
How Much VRAM Does LFM2.5 2.6B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 1.6 GB | 2.3 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 128K context · ~417 tok/s | |
| 1.6 GB | 2.3 GB | 3-bit small — 3-bit small quantization | Fits: 128K context · ~407 tok/s | |
| 1.7 GB | 2.4 GB | 3-bit medium — 3-bit medium quantization | Fits: 128K context · ~377 tok/s | |
| 1.8 GB | 2.5 GB | 4-bit · legacy — 4-bit legacy quantization | Fits: 128K context · ~370 tok/s | |
| 2.0 GB | 2.7 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 128K context · ~321 tok/s | |
| 2.4 GB | 3 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 128K context · ~279 tok/s | |
| 2.6 GB | 3.3 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 128K context · ~247 tok/s | |
| 3.1 GB | 3.8 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 128K context · ~210 tok/s |
Selected Q4_K_M
4-bit medium quantization2.3GBat 16K context
- Weights
- 1.6 GB
- KV cache
- 0.4 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 21.7 GB free · ~321 tok/s
≥50 tok/s needs ≥157 GB/s · RTX 5090 ~571, RTX 3090 Ti ~321
Which Hardware Can Run LFM2.5 2.6B?
Q4_K_M · 2.0 GBLFM2.5 2.6B (Q4_K_M) requires 2.0 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 3+ GB is recommended. Using the full 131K context window can add up to 3.1 GB, bringing total usage to 5.1 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run LFM2.5 2.6B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 5.7 GB free · up to 128K ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~242Simulate output on RTX 3080RTX 2080 Ti~196Simulate output on RTX 2080 TiRTX 3070 Ti~194Simulate output on RTX 3070 TiM3*8 GB~35Simulate output on iMac 24" M3 (8 GB)Orin Nano~33Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~23Simulate output on iPhone 1720 GB
fits · 17.7 GB free · up to 128K ctx
1 GPU
RX 7900 XT~235Simulate output on RX 7900 XT32–40 GB
fits · 29.7 GB free · up to 128K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~571Simulate output on RTX 5090M5 Max36 GB~158Simulate output on Mac Studio M5 Max (36 GB)M532 GB~53Simulate output on MacBook Air 13" M5 (32 GB)Orin~65Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~43Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to LFM2.5 2.6B
LFM2.5 family · 14 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| MiniCPM5 2BOpenBMB | — | 2.5 GB | ~345 | same size |
| Granite 3.1 2B InstructIBM | — | 3.2 GB | ~329 | same size |
| Gemma 2 2B ITGoogle | — | 2.5 GB | ~379 | same size |
| Phi 2Microsoft | — | 2.6 GB | ~248 | same size |
| Ouro 2.6B ThinkingByteDance | — | 12.1 GB | ~102 | same size |
| Nemotron Flash 3BNVIDIA | — | 7.6 GB | ~109 | same size |
Fine-tunes & variants(5)
Specifications
- Publisher
- Liquid AI
- Family
- LFM2.5
- Parameters
- 2.7B
- Architecture
- Lfm2ForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 128,000
- Release Date
- 2026-07-28
- License
- Other
Frequently Asked Questions
How much VRAM does LFM2.5 2.6B need?
LFM2.5 2.6B requires 2.0 GB of VRAM at Q4_K_M, or 5.8 GB at BF16. Full 131K context adds up to 3.1 GB (5.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 2.7B × 4.8 bits ÷ 8 = 1.6 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 3.5 GB (at full 131K context)
VRAM usage by quantization
What's the best quantization for LFM2.5 2.6B?
For LFM2.5 2.6B, Q4_K_M (2.0 GB) offers the best balance of quality and VRAM usage. Q4_K_L (2.1 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 1.2 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run LFM2.5 2.6B on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run LFM2.5 2.6B. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 2.0 GB of usable unified memory (RAM minus macOS overhead).
Can I run LFM2.5 2.6B locally?
Yes — LFM2.5 2.6B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is LFM2.5 2.6B?
At Q4_K_M, LFM2.5 2.6B can reach ~2353 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~321 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 2.04 × 0.65 = ~2549 tok/s
Estimated speed at Q4_K_M (2.0 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of LFM2.5 2.6B?
At Q4_K_M, the download is about 1.62 GB. The full-precision BF16 version is 5.39 GB. The smallest option (IQ2_XXS) is 0.74 GB.
Which GPUs can run LFM2.5 2.6B?
52 consumer GPUs can run LFM2.5 2.6B at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
Which devices can run LFM2.5 2.6B?
163 devices with unified memory can run LFM2.5 2.6B at Q4_K_M (2.0 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.