Model benchmarks
Ornith-1.5-35B-A3B-APEX-MTP-I-Compact local LLM performance
As of September 2026, Ornith-1.5-35B-A3B-APEX-MTP-I-Compact runs at up to 40.0 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
35B
Peak speed
40.0 tok/s
Average speed
31.5 tok/s
Avg PP
603.5 tok/s
Min memory
n/a
Max context
131.072 tokens
Avg output / run
26.586 tokens
Avg runtime / run
13m 57s
Avg quality
83.7
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 | 81.9 | 83.7 | 85.4 |
| Agent Workflow | 85.5 | 88.1 | 90.7 |
| Code Generation | 79.4 | 79.6 | 79.8 |
| Role Play & Narrative | 79.9 | 84.6 | 89.3 |
| Research & Analysis | 82.3 | 82.4 | 82.4 |
Performance by hardware and tool
Every hardware/tool/quantization combination Ornith-1.5-35B-A3B-APEX-MTP-I-Compact 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 |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 9050 / 9060 XT | LM Studio | — | 40.0 tok/s | 40.0 tok/s | 16.2 GB | 131.072 tokens | 85.6 | 1 |
| CPU only | vLLM | — | 23.0 tok/s | 23.0 tok/s | n/a | 98.304 tokens | 81.7 | 1 |
Benchmark runs
All 2 Ornith-1.5-35B-A3B-APEX-MTP-I-Compact runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Ornith-1.5-35B-A3B-APEX-MTP-I-Compact good for coding?
- In our benchmarks, Ornith-1.5-35B-A3B-APEX-MTP-I-Compact scores 79.6/100 for coding. It runs at about 31.5 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 Ornith-1.5-35B-A3B-APEX-MTP-I-Compact good for agentic (tool-using) tasks?
- In our benchmarks, Ornith-1.5-35B-A3B-APEX-MTP-I-Compact scores 88.1/100 for agentic workflows. It runs at about 31.5 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 Ornith-1.5-35B-A3B-APEX-MTP-I-Compact for local inference?
- Across 2 community benchmark runs, Ornith-1.5-35B-A3B-APEX-MTP-I-Compact reaches up to 40.0 tok/s and averages 31.5 tok/s, with the fastest results on AMD Radeon RX 9050 / 9060 XT.
- How much memory does Ornith-1.5-35B-A3B-APEX-MTP-I-Compact need?
- The leanest observed configuration used about n/a of memory.
- Which tools have been used to run Ornith-1.5-35B-A3B-APEX-MTP-I-Compact?
- Benchmarks were submitted using LM Studio, vLLM. Results are community-contributed and updated as new runs arrive.