Model benchmarks
Tiel-Coder-35B-A3B-Q4_K_S local LLM performance
As of September 2026, Tiel-Coder-35B-A3B-Q4_K_S runs at up to 166.6 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
35B
Peak speed
166.6 tok/s
Average speed
165.0 tok/s
Min memory
17.1 GB
Max context
65,536 tokens
Avg output / run
27,382 tokens
Avg runtime / run
2m 43s
Avg quality
84.6
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 | 83.9 | 84.6 | 85.1 |
| Agent Workflow | 85.9 | 89.3 | 92.3 |
| Code Generation | 75.3 | 76.9 | 79.4 |
| Role Play & Narrative | 82.9 | 87.3 | 91.4 |
| Research & Analysis | 83.0 | 84.9 | 87.1 |
Performance by hardware and tool
Every hardware/tool/quantization combination Tiel-Coder-35B-A3B-Q4_K_S 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 7900 XT/7900 XTX/7900 GRE/7900M | llama.cpp | Q4_K_S | 166.6 tok/s | 165.0 tok/s | 17.1 GB | 65,536 tokens | 84.6 | 3 |
Frequently asked questions
- How fast is Tiel-Coder-35B-A3B-Q4_K_S for local inference?
- Across 3 community benchmark runs, Tiel-Coder-35B-A3B-Q4_K_S reaches up to 166.6 tok/s and averages 165.0 tok/s, with the fastest results on AMD Radeon RX 7900 XT/7900 XTX/7900 GRE/7900M.
- How much memory does Tiel-Coder-35B-A3B-Q4_K_S need?
- The leanest observed configuration used about 17.1 GB of memory (quantizations tested: Q4_K_S).
- Which tools have been used to run Tiel-Coder-35B-A3B-Q4_K_S?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.