Model benchmarks
Ornith-35B-A3B-GSQ local LLM performance
As of October 2026, Ornith-35B-A3B-GSQ runs at up to 113.5 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
35B
Peak speed
113.5 tok/s
Average speed
113.3 tok/s
Avg PP
1497.6 tok/s
Min memory
13.6 GB
Max context
65,536 tokens
Avg output / run
26,297 tokens
Avg runtime / run
3m 58s
Avg quality
74.9
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 | 65.9 | 74.9 | 83.8 |
| Agent Workflow | 17.6 | 50.3 | 82.9 |
| Code Generation | 76.1 | 79.3 | 82.4 |
| Role Play & Narrative | 83.8 | 84.4 | 85.0 |
| Research & Analysis | 84.8 | 85.6 | 86.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination Ornith-35B-A3B-GSQ has been benchmarked on, ranked by peak token generation speed. Last updated October 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7800 XT | llama.cpp | — | 113.5 tok/s | 113.3 tok/s | 13.6 GB | 65,536 tokens | 74.9 | 2 |
Benchmark runs
All 2 Ornith-35B-A3B-GSQ runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Ornith-35B-A3B-GSQ good for agentic (tool-using) tasks?
- In our benchmarks, Ornith-35B-A3B-GSQ scores 50.3/100 for agentic workflows. It runs at about 113.3 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-MLX-oQ4e-MTP is faster (~151.8 tok/s) and still scores well for agentic workflows (83.0/100).
- How fast is Ornith-35B-A3B-GSQ for local inference?
- Across 2 community benchmark runs, Ornith-35B-A3B-GSQ reaches up to 113.5 tok/s and averages 113.3 tok/s, with the fastest results on AMD Radeon RX 7800 XT.
- How much memory does Ornith-35B-A3B-GSQ need?
- The leanest observed configuration used about 13.6 GB of memory.
- Which tools have been used to run Ornith-35B-A3B-GSQ?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.