Model benchmarks
granite-4.2-30b-MLX-8bit local LLM performance
As of September 2026, granite-4.2-30b-MLX-8bit runs at up to 16.1 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
30B
Peak speed
16.1 tok/s
Average speed
16.0 tok/s
Avg PP
608.5 tok/s
Min memory
31.0 GB
Max context
131.072 tokens
Avg output / run
35.800 tokens
Avg runtime / run
38m 17s
Avg quality
72.1
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 | 71.1 | 72.1 | 73.2 |
| Agent Workflow | 69.0 | 73.1 | 77.1 |
| Code Generation | 49.4 | 51.5 | 53.5 |
| Role Play & Narrative | 88.5 | 89.3 | 90.2 |
| Research & Analysis | 73.4 | 74.7 | 76.0 |
Performance by hardware and tool
Every hardware/tool/quantization combination granite-4.2-30b-MLX-8bit 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 M5 Max | oMLX | 8bit | 16.1 tok/s | 16.0 tok/s | 31.0 GB | 131.072 tokens | 72.1 | 2 |
Benchmark runs
All 2 granite-4.2-30b-MLX-8bit runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is granite-4.2-30b-MLX-8bit good for coding?
- In our benchmarks, granite-4.2-30b-MLX-8bit scores 51.5/100 for coding. It runs at about 16.0 tok/s, so if you want more speed, IsValorum/Nex-N2.5-mini-APEX-I-MiniPlus-V2.1-GGUF:1 is faster (~25.7 tok/s) and still scores well for coding (85.0/100).
- Is granite-4.2-30b-MLX-8bit good for agentic (tool-using) tasks?
- In our benchmarks, granite-4.2-30b-MLX-8bit scores 73.1/100 for agentic workflows. It runs at about 16.0 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 granite-4.2-30b-MLX-8bit for local inference?
- Across 2 community benchmark runs, granite-4.2-30b-MLX-8bit reaches up to 16.1 tok/s and averages 16.0 tok/s, with the fastest results on Apple M5 Max.
- How much memory does granite-4.2-30b-MLX-8bit need?
- The leanest observed configuration used about 31.0 GB of memory (quantizations tested: 8bit).
- Which tools have been used to run granite-4.2-30b-MLX-8bit?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.