Model benchmarks
IFM--K2-Horizon-MoVA-36B-A4B local LLM performance
As of September 2026, IFM--K2-Horizon-MoVA-36B-A4B runs at up to 31.9 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
36B
Peak speed
31.9 tok/s
Average speed
31.8 tok/s
Avg prefill
338.4 tok/s
Min memory
71.6 GB
Max context
131,072 tokens
Avg output / run
51,947 tokens
Avg runtime / run
29m 48s
Avg quality
73.5
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 | 66.5 | 73.5 | 80.5 |
| Agent Workflow | 87.1 | 89.0 | 90.8 |
| Code Generation | 13.1 | 41.2 | 69.3 |
| Role Play & Narrative | 81.7 | 83.9 | 86.2 |
| Research & Analysis | 79.4 | 79.9 | 80.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination IFM--K2-Horizon-MoVA-36B-A4B has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Apple M4 Max | oMLX | — | 31.9 tok/s | 31.8 tok/s | 71.6 GB | 131,072 tokens | 73.5 | 2 |
Frequently asked questions
- Is IFM--K2-Horizon-MoVA-36B-A4B good for coding?
- In our benchmarks, IFM--K2-Horizon-MoVA-36B-A4B scores 41.2/100 for coding. It runs at about 31.8 tok/s, so if you want more speed, unsloth/Qwen3.8-27B-GGUF:IQ3_S is faster (~104.3 tok/s) and still scores well for coding (83.0/100).
- Is IFM--K2-Horizon-MoVA-36B-A4B good for agentic (tool-using) tasks?
- In our benchmarks, IFM--K2-Horizon-MoVA-36B-A4B scores 89.0/100 for agentic workflows. It runs at about 31.8 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 IFM--K2-Horizon-MoVA-36B-A4B for local inference?
- Across 2 community benchmark runs, IFM--K2-Horizon-MoVA-36B-A4B reaches up to 31.9 tok/s and averages 31.8 tok/s, with the fastest results on Apple M4 Max.
- How much memory does IFM--K2-Horizon-MoVA-36B-A4B need?
- The leanest observed configuration used about 71.6 GB of memory.
- Which tools have been used to run IFM--K2-Horizon-MoVA-36B-A4B?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.