Model benchmarks
ornith-1.5-35b-a3b-crack-q4_k_m local LLM performance
As of November 2026, ornith-1.5-35b-a3b-crack-q4_k_m runs at up to 106.5 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
35B
Peak speed
106.5 tok/s
Average speed
104.0 tok/s
Avg PP
406.7 tok/s
Min memory
11.9 GB
Max context
65,536 tokens
Avg output / run
33,997 tokens
Avg runtime / run
5m 32s
Avg quality
73.3
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 | 66.1 | 73.3 | 80.4 |
| Agent Workflow | 16.2 | 45.1 | 73.9 |
| Code Generation | 74.8 | 76.6 | 78.4 |
| Role Play & Narrative | 86.1 | 87.5 | 89.0 |
| Research & Analysis | 83.4 | 83.9 | 84.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination ornith-1.5-35b-a3b-crack-q4_k_m 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 4070 | llama.cpp | Q4_K_M | 106.5 tok/s | 104.0 tok/s | 11.9 GB | 65,536 tokens | 73.3 | 2 |
Benchmark runs
All 2 ornith-1.5-35b-a3b-crack-q4_k_m runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is ornith-1.5-35b-a3b-crack-q4_k_m good for agentic (tool-using) tasks?
- In our benchmarks, ornith-1.5-35b-a3b-crack-q4_k_m scores 45.1/100 for agentic workflows. It runs at about 104.0 tok/s, so if you want more speed, qwen3.8-flash-next-q2_0 is faster (~127.9 tok/s) and still scores well for agentic workflows (85.6/100).
- How fast is ornith-1.5-35b-a3b-crack-q4_k_m for local inference?
- Across 2 community benchmark runs, ornith-1.5-35b-a3b-crack-q4_k_m reaches up to 106.5 tok/s and averages 104.0 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.
- How much memory does ornith-1.5-35b-a3b-crack-q4_k_m need?
- The leanest observed configuration used about 11.9 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run ornith-1.5-35b-a3b-crack-q4_k_m?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.