Model benchmarks
Tiel-Coder-35B-A3B-MLX-oQ4e-MTP local LLM performance
As of August 2026, Tiel-Coder-35B-A3B-MLX-oQ4e-MTP runs at up to 126.6 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).
oMLX
Model size
35B
Peak speed
126.6 tok/s
Average speed
123.6 tok/s
Min memory
21.0 GB
Max context
262,144 tokens
Avg output / run
25,968 tokens
Avg runtime / run
3m 37s
Avg quality
81.8
Benchmark runs
4
GPUs tested
1
Performance by hardware and tool
Every hardware/tool/quantization combination Tiel-Coder-35B-A3B-MLX-oQ4e-MTP has been benchmarked on, ranked by peak token generation speed. Last updated August 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Apple M5 Max | oMLX | — | 126.6 tok/s | 123.6 tok/s | 21.0 GB | 262,144 tokens | 81.8 | 4 |
Frequently asked questions
- How fast is Tiel-Coder-35B-A3B-MLX-oQ4e-MTP for local inference?
- Across 4 community benchmark runs, Tiel-Coder-35B-A3B-MLX-oQ4e-MTP reaches up to 126.6 tok/s and averages 123.6 tok/s, with the fastest results on Apple M5 Max.
- How much memory does Tiel-Coder-35B-A3B-MLX-oQ4e-MTP need?
- The leanest observed configuration used about 21.0 GB of memory.
- Which tools have been used to run Tiel-Coder-35B-A3B-MLX-oQ4e-MTP?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.