Model benchmarks
Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL local LLM performance
As of September 2026, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL runs at up to 50.2 tok/s for local inference (best of 5 community benchmark runs across 2 GPUs).
Model size
35B
Peak speed
50.2 tok/s
Average speed
37.4 tok/s
Avg PP
614.4 tok/s
Min memory
n/a
Max context
65,536 tokens
Avg output / run
20,259 tokens
Avg runtime / run
9m 16s
Avg quality
78.5
Benchmark runs
5
GPUs tested
2
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 5 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 | 67.4 | 78.5 | 84.9 |
| Agent Workflow | 79.6 | 87.4 | 93.5 |
| Code Generation | 33.4 | 66.6 | 80.3 |
| Role Play & Narrative | 71.5 | 81.9 | 89.2 |
| Research & Analysis | 66.2 | 78.0 | 84.4 |
Performance by hardware and tool
Every hardware/tool/quantization combination Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL 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 8050S | llama.cpp | Q6_K | 50.2 tok/s | 43.5 tok/s | n/a | 65,536 tokens | 76.9 | 3 |
| Tesla P100-PCIE-16GB | llama.cpp | Q6_K | 28.2 tok/s | 28.2 tok/s | n/a | 65,536 tokens | 80.9 | 2 |
Benchmark runs
All 5 Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL good for coding?
- In our benchmarks, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL scores 66.6/100 for coding. It runs at about 37.4 tok/s, so if you want more speed, local-qwen is faster (~60.2 tok/s) and still scores well for coding (84.8/100).
- Is Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL good for agentic (tool-using) tasks?
- In our benchmarks, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL scores 87.4/100 for agentic workflows. It runs at about 37.4 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 Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL for local inference?
- Across 5 community benchmark runs, Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL reaches up to 50.2 tok/s and averages 37.4 tok/s, with the fastest results on AMD Radeon 8050S.
- Which tools have been used to run Cyber-Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.