Model benchmarks
LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 local LLM performance
As of November 2026, LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 runs at up to 29.1 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
24B
Peak speed
29.1 tok/s
Average speed
26.3 tok/s
Avg PP
228.8 tok/s
Min memory
25.6 GB
Max context
131,072 tokens
Avg output / run
31,772 tokens
Avg runtime / run
23m 34s
Avg quality
68.7
Benchmark runs
2
GPUs tested
1
Quality by task
Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 2 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.
| Task | P5 (low) | Avg | P95 (high) |
|---|---|---|---|
| Overall | 59.5 | 68.7 | 77.8 |
| Agent Workflow | 22.6 | 52.7 | 82.9 |
| Code Generation | 67.4 | 68.8 | 70.1 |
| Role Play & Narrative | 78.7 | 79.1 | 79.6 |
| Research & Analysis | 69.2 | 74.0 | 78.7 |
Performance by hardware and tool
Every hardware/tool/quantization combination LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 2060 | llama.cpp | Q4_0 | 29.1 tok/s | 26.3 tok/s | 25.6 GB | 131,072 tokens | 68.7 | 2 |
Benchmark runs
All 2 LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 good for coding?
- In our benchmarks, LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 scores 68.8/100 for coding. It runs at about 26.3 tok/s, so if you want more speed, Qwen3.8-27B-Cold-Fusion-GAIN-V1.1-NM-DAU-NEO-MAX-MTP-GGUF-Q6_K is faster (~34.2 tok/s) and still scores well for coding (85.9/100).
- Is LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 good for agentic (tool-using) tasks?
- In our benchmarks, LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 scores 52.7/100 for agentic workflows. It runs at about 26.3 tok/s, so if you want more speed, granite-4.2-8b-Q4_K_L is faster (~71.6 tok/s) and still scores well for agentic workflows (92.9/100).
- How fast is LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 for local inference?
- Across 2 community benchmark runs, LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 reaches up to 29.1 tok/s and averages 26.3 tok/s, with the fastest results on NVIDIA GeForce RTX 2060.
- How much memory does LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0 need?
- The leanest observed configuration used about 25.6 GB of memory (quantizations tested: Q4_0).
- Which tools have been used to run LiquidAI/LFM2.5-24B-A2B-GGUF:Q4_0?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.