Model benchmarks
Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 local LLM performance
As of October 2026, Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 runs at up to 44.7 tok/s for local inference (best of 13 community benchmark runs across 2 GPUs).
Model size
35B
Peak speed
44.7 tok/s
Average speed
23.0 tok/s
Avg PP
256.9 tok/s
Min memory
16.3 GB
Max context
200,000 tokens
Avg output / run
23,756 tokens
Avg runtime / run
22m 39s
Avg quality
80.8
Benchmark runs
13
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 13 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 | 75.1 | 80.8 | 84.2 |
| Agent Workflow | 77.2 | 82.7 | 88.1 |
| Code Generation | 68.3 | 79.2 | 85.2 |
| Role Play & Narrative | 53.9 | 77.3 | 87.3 |
| Research & Analysis | 79.8 | 83.8 | 88.1 |
Performance by hardware and tool
Every hardware/tool/quantization combination Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 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 7900 XTX | LM Studio | — | 44.7 tok/s | 44.7 tok/s | 21.6 GB | 80,000 tokens | 82.8 | 1 |
| CPU only | LM Studio | — | 31.5 tok/s | 27.0 tok/s | 16.3 GB | 65,536 tokens | 78.1 | 4 |
| Intel Graphics | llama.cpp | — | 21.9 tok/s | 18.8 tok/s | n/a | 200,000 tokens | 81.6 | 6 |
| Intel Graphics | llama.cpp:vulkan | — | 20.3 tok/s | 20.3 tok/s | n/a | 200,000 tokens | 82.6 | 1 |
| Intel Graphics | llama.cpp:sycl | — | 12.6 tok/s | 12.6 tok/s | n/a | 200,000 tokens | 82.5 | 1 |
Benchmark runs
All 13 Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 good for coding?
- In our benchmarks, Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 scores 79.2/100 for coding. It runs at about 23.0 tok/s, so if you want more speed, byteshape/Qwen3.8-27B-GGUF:IQ4_XS-ASCII-Condensed is faster (~42.0 tok/s) and still scores well for coding (86.3/100).
- Is Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 good for agentic (tool-using) tasks?
- In our benchmarks, Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 scores 82.7/100 for agentic workflows. It runs at about 23.0 tok/s, so if you want more speed, muse-glimmer:latest is faster (~33.3 tok/s) and still scores well for agentic workflows (91.6/100).
- How fast is Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 for local inference?
- Across 13 community benchmark runs, Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 reaches up to 44.7 tok/s and averages 23.0 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
- How much memory does Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1 need?
- The leanest observed configuration used about 16.3 GB of memory.
- Which tools have been used to run Cyber-Tiel-Coder-35B-A3B.APEX-I-MiniPlus-V2.1?
- Benchmarks were submitted using LM Studio, llama.cpp, llama.cpp:sycl, llama.cpp:vulkan. Results are community-contributed and updated as new runs arrive.