Model benchmarks
Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS local LLM performance
As of September 2026, Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS runs at up to 88.5 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
35B
Peak speed
88.5 tok/s
Average speed
87.8 tok/s
Avg PP
1003.9 tok/s
Min memory
17.1 GB
Max context
262.144 tokens
Avg output / run
20.510 tokens
Avg runtime / run
3m 59s
Avg quality
83.9
Benchmark runs
3
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 3 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 | 82.3 | 83.9 | 86.6 |
| Agent Workflow | 77.0 | 82.5 | 86.5 |
| Code Generation | 69.3 | 75.7 | 80.7 |
| Role Play & Narrative | 92.0 | 92.7 | 93.3 |
| Research & Analysis | 82.1 | 84.8 | 88.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS 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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 5060 | llama.cpp | — | 88.5 tok/s | 87.8 tok/s | 17.1 GB | 262.144 tokens | 83.9 | 3 |
Benchmark runs
All 3 Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS good for coding?
- In our benchmarks, Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS scores 75.7/100 for coding. It runs at about 87.8 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 Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS good for agentic (tool-using) tasks?
- In our benchmarks, Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS scores 82.5/100 for agentic workflows. It runs at about 87.8 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~165.0 tok/s) and still scores well for agentic workflows (89.3/100).
- How fast is Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS for local inference?
- Across 3 community benchmark runs, Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS reaches up to 88.5 tok/s and averages 87.8 tok/s, with the fastest results on NVIDIA GeForce RTX 5060.
- How much memory does Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS need?
- The leanest observed configuration used about 17.1 GB of memory.
- Which tools have been used to run Unsloth-Ornith-1.5-35B-A3B-UD-IQ3_XXS?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.