Model benchmarks
ornith-ai/Ornith-1.5-122B-MLX-4bit local LLM performance
As of November 2026, ornith-ai/Ornith-1.5-122B-MLX-4bit runs at up to 115.6 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
122B
Peak speed
115.6 tok/s
Average speed
113.4 tok/s
Avg PP
1816.4 tok/s
Min memory
68.5 GB
Max context
65,536 tokens
Avg output / run
16,669 tokens
Avg runtime / run
2m 24s
Avg quality
84.4
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 | 84.2 | 84.4 | 84.5 |
| Agent Workflow | 81.7 | 83.4 | 85.0 |
| Code Generation | 75.1 | 77.4 | 79.8 |
| Role Play & Narrative | 88.7 | 89.1 | 89.5 |
| Research & Analysis | 86.7 | 87.6 | 88.5 |
Performance by hardware and tool
Every hardware/tool/quantization combination ornith-ai/Ornith-1.5-122B-MLX-4bit 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 |
|---|---|---|---|---|---|---|---|---|
| Apple M5 Max | splash | — | 115.6 tok/s | 113.4 tok/s | 68.5 GB | 65,536 tokens | 84.4 | 2 |
Benchmark runs
All 2 ornith-ai/Ornith-1.5-122B-MLX-4bit runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is ornith-ai/Ornith-1.5-122B-MLX-4bit good for coding?
- In our benchmarks, ornith-ai/Ornith-1.5-122B-MLX-4bit scores 77.4/100 for coding. It runs at about 113.4 tok/s, so if you want more speed, Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well for coding (82.2/100).
- Is ornith-ai/Ornith-1.5-122B-MLX-4bit good for agentic (tool-using) tasks?
- In our benchmarks, ornith-ai/Ornith-1.5-122B-MLX-4bit scores 83.4/100 for agentic workflows. It runs at about 113.4 tok/s, so if you want more speed, swift-1.5-qwen3.8-flash-next-iq3_s is faster (~259.0 tok/s) and still scores well for agentic workflows (88.9/100).
- How fast is ornith-ai/Ornith-1.5-122B-MLX-4bit for local inference?
- Across 2 community benchmark runs, ornith-ai/Ornith-1.5-122B-MLX-4bit reaches up to 115.6 tok/s and averages 113.4 tok/s, with the fastest results on Apple M5 Max.
- How much memory does ornith-ai/Ornith-1.5-122B-MLX-4bit need?
- The leanest observed configuration used about 68.5 GB of memory.
- Which tools have been used to run ornith-ai/Ornith-1.5-122B-MLX-4bit?
- Benchmarks were submitted using splash. Results are community-contributed and updated as new runs arrive.