Total runs:
4.1K
Run Growth:
-365
Growth Rate:
-8.78%
Sponsored by
Vmake Labs Video Enhancer
Anbeeld / LFM2.5-1.2B-Instruct-DSpark-GGUF
Model's Last Updated: September 07 2026
text-generation
Introduction of LFM2.5-1.2B-Instruct-DSpark-GGUF
Model Details of LFM2.5-1.2B-Instruct-DSpark-GGUF
LFM2.5-1.2B-Instruct DSpark GGUF
GGUF quantizations of LiquidAI DSpark draft model for LFM2.5-1.2B-Instruct .
Use with BeeLlama.cpp , a llama.cpp fork with advanced quantization features.
LFM2.5-1.2B-Instruct-DSpark
LFM2.5-DSpark is a family of speculative-decoding draft models that adapt DSpark for the LFM2.5 architecture. They allow LFM2.5 models to run faster without degrading quality.
This is a drafter for
LiquidAI/LFM2.5-1.2B-Instruct
.
In SGLang, decoding runs about 2× faster. It also runs on-device on Apple silicon through the Metal backend.
Find more information about LFM2.5-DSpark in our blog post .
🗒️ Model Details
LFM2.5-1.2B-Instruct-DSpark is a DSpark speculative-decoding draft model with the following features:
-
Target model
:
LiquidAI/LFM2.5-1.2B-Instruct - Draft parameters : 295.7M (BF16)
-
Backbone
: 5 full attention layers,
hidden_size=2048,intermediate_size=6144with SiLU/SwiGLU, GQA withnum_attention_heads=32/num_key_value_heads=8,head_dim=64 - Extra heads : Markov head (rank 256) + confidence head
- Block size : 9
- Vocabulary : 65,536
Other models in the LFM2.5-DSpark family:
| Drafter | Target |
|---|---|
| LFM2.5-1.2B-Instruct-DSpark | LFM2.5-1.2B-Instruct |
| LFM2.5-8B-A1B-DSpark | LFM2.5-8B-A1B |
| LFM2.5-2.6B-DSpark | LFM2.5-2.6B |
📊 Performance
Benchmarks
Speculative decoding is
exact
: the target verifies every proposed token, so the generated
text is what the target would have produced on its own. See
LiquidAI/LFM2.5-1.2B-Instruct
for performance benchmarks.
Acceptance
Mean accepted tokens per decoding step, by benchmark (1×H100, batch size 1, greedy decoding). Higher means more of the draft's proposed block is accepted per target forward pass, so decoding is faster (at block size 9, the ceiling is 10).
| Benchmark | Accepted tokens / step |
|---|---|
| MATH-500 | 5.78 |
| GSM8K | 4.25 |
| HumanEval | 5.51 |
| MBPP | 5.41 |
| MT-Bench | 3.11 |
| Mean | 4.81 |
On-device and GPU Inference
| Dataset | Acceptance (of 10) | Speedup on H100 | Speedup on M4 Max |
|---|---|---|---|
| MATH500 | 6.02 |
2.56x
668 → 1712 tok/s |
2.62x
140 → 366 tok/s |
| HumanEval | 5.31 |
2.26x
664 → 1499 tok/s |
2.87x
136 → 389 tok/s |
| MBPP | 5.52 |
2.37x
667 → 1578 tok/s |
2.74x
137 → 375 tok/s |
| GSM8K | 4.34 |
1.67x
624 → 1041 tok/s |
2.73x
140 → 381 tok/s |
| MT-Bench | 3.90 |
1.66x
657 → 1091 tok/s |
1.72x
137 → 237 tok/s |
| Mean | 5.02 |
2.10x
656 → 1384 tok/s |
2.54x
138 → 350 tok/s |
🏃 How to run (SGLang)
Requires a build of SGLang with DSpark support for LFM2 targets ( PR #31041 ). Launch the target with the drafter attached:
python -m sglang.launch_server \
--model-path LiquidAI/LFM2.5-1.2B-Instruct \
--speculative-algorithm DSPARK \
--speculative-draft-model-path LiquidAI/LFM2.5-1.2B-Instruct-DSpark \
--speculative-draft-attention-backend flashinfer \
--disable-radix-cache --mem-fraction-static 0.75 --port 30000
Then query the OpenAI-compatible endpoint at
http://localhost:30000/v1
. The block size is read
from the draft's
config.json
; the baseline is the same command without the three
--speculative-*
flags.
📬 Contact
- Got questions or want to connect? Join our Discord community
- If you are interested in custom solutions with edge deployment, please contact our sales team .
Citation
@article{liquidAI202626B,
author = {Liquid AI},
title = {LFM2.5-2.6B: Agents Everywhere},
journal = {Liquid AI Blog},
year = {2026},
note = {www.liquid.ai/blog/lfm2-5-2-6b},
}
@article{liquidAI2026dspark,
author = {Liquid AI},
title = {LFM2.5-DSpark: Up to 3.2x Faster Inference from H100 to MacBook},
journal = {Liquid AI Blog},
year = {2026},
note = {www.liquid.ai/blog/lfm2.5-dspark},
}
Runs of Anbeeld LFM2.5-1.2B-Instruct-DSpark-GGUF on huggingface.co
200
Total runs
0
24-hour runs
-18
3-day runs
-16
7-day runs
10
30-day runs
More Information About LFM2.5-1.2B-Instruct-DSpark-GGUF huggingface.co Model
Url of LFM2.5-1.2B-Instruct-DSpark-GGUF
LFM2.5-1.2B-Instruct-DSpark-GGUF huggingface.co Url
Provider of LFM2.5-1.2B-Instruct-DSpark-GGUF huggingface.co
Other API from Anbeeld
Updated:August 30 2026
Total runs:
3.7K
Run Growth:
-3.1K
Growth Rate:
-81.89%
Total runs:
3.0K
Run Growth:
-2.0K
Growth Rate:
-66.58%
Total runs:
3.0K
Run Growth:
1.5K
Growth Rate:
50.63%
Total runs:
2.9K
Run Growth:
-2.2K
Growth Rate:
-76.32%
Total runs:
2.8K
Run Growth:
-1.9K
Growth Rate:
-77.95%
Total runs:
2.0K
Run Growth:
-2.0K
Growth Rate:
-102.35%
Total runs:
1.7K
Run Growth:
-513
Growth Rate:
-28.97%
Total runs:
1.0K
Run Growth:
-427
Growth Rate:
-42.83%
Total runs:
965
Run Growth:
-329
Growth Rate:
-37.77%
Total runs:
932
Run Growth:
408
Growth Rate:
43.64%
Total runs:
914
Run Growth:
583
Growth Rate:
73.15%
Total runs:
905
Run Growth:
319
Growth Rate:
34.90%
Total runs:
871
Run Growth:
248
Growth Rate:
26.99%
Total runs:
826
Run Growth:
-1.4K
Growth Rate:
-181.23%
Total runs:
688
Run Growth:
286
Growth Rate:
40.80%
Total runs:
649
Run Growth:
79
Growth Rate:
12.31%
Total runs:
638
Run Growth:
74
Growth Rate:
11.64%
Total runs:
612
Run Growth:
207
Growth Rate:
33.66%
Total runs:
611
Run Growth:
-157
Growth Rate:
-25.70%
Total runs:
608
Run Growth:
-108
Growth Rate:
-18.15%
Total runs:
536
Run Growth:
22
Growth Rate:
3.99%
Total runs:
515
Run Growth:
97
Growth Rate:
18.98%
Total runs:
502
Run Growth:
85
Growth Rate:
16.63%
Total runs:
457
Run Growth:
-30
Growth Rate:
-6.55%
Total runs:
320
Run Growth:
-153
Growth Rate:
-45.67%
Total runs:
309
Run Growth:
-495
Growth Rate:
-160.71%
Total runs:
260
Run Growth:
-29
Growth Rate:
-10.74%
Total runs:
257
Run Growth:
-293
Growth Rate:
-120.58%
Total runs:
252
Run Growth:
-150
Growth Rate:
-59.76%
Total runs:
246
Run Growth:
-63
Growth Rate:
-25.20%
Total runs:
246
Run Growth:
-332
Growth Rate:
-123.42%
Total runs:
225
Run Growth:
-192
Growth Rate:
-106.08%
Total runs:
220
Run Growth:
-77
Growth Rate:
-36.15%
Total runs:
210
Run Growth:
-339
Growth Rate:
-161.43%
Total runs:
181
Run Growth:
-286
Growth Rate:
-156.28%
Total runs:
180
Run Growth:
-226
Growth Rate:
-121.51%
Total runs:
173
Run Growth:
-97
Growth Rate:
-58.43%
Total runs:
161
Run Growth:
-368
Growth Rate:
-213.95%
Total runs:
159
Run Growth:
-123
Growth Rate:
-80.92%
Total runs:
141
Run Growth:
-272
Growth Rate:
-214.17%
Total runs:
132
Run Growth:
-578
Growth Rate:
-437.88%
Total runs:
0
Run Growth:
0
Growth Rate:
0.00%
Total runs:
0
Run Growth:
0
Growth Rate:
0.00%
Total runs:
0
Run Growth:
0
Growth Rate:
0.00%
Total runs:
0
Run Growth:
0
Growth Rate:
0.00%
Total runs:
0
Run Growth:
0
Growth Rate:
0.00%