Model benchmarks
Qwen3-Coder-Next_t1p95k40m0 local LLM performance
As of November 2026, Qwen3-Coder-Next_t1p95k40m0 runs at up to 81.9 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
2.5B
Peak speed
81.9 tok/s
Average speed
81.5 tok/s
Avg PP
231.3 tok/s
Min memory
56.8 GB
Max context
32,768 tokens
Avg output / run
8,529 tokens
Avg runtime / run
1m 54s
Avg quality
78.5
Benchmark runs
2
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 2 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 | 77.5 | 78.5 | 79.5 |
| Agent Workflow | 89.4 | 89.4 | 89.5 |
| Code Generation | 68.1 | 70.4 | 72.6 |
| Role Play & Narrative | 76.9 | 85.0 | 93.1 |
| Research & Analysis | 67.5 | 69.3 | 71.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination Qwen3-Coder-Next_t1p95k40m0 has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Tesla V100-PCIE-32GB | llama.cpp | — | 81.9 tok/s | 81.5 tok/s | 56.8 GB | 32,768 tokens | 78.5 | 2 |
Benchmark runs
All 2 Qwen3-Coder-Next_t1p95k40m0 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Qwen3-Coder-Next_t1p95k40m0 good for coding?
- In our benchmarks, Qwen3-Coder-Next_t1p95k40m0 scores 70.4/100 for coding. It runs at about 81.5 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 Qwen3-Coder-Next_t1p95k40m0 good for agentic (tool-using) tasks?
- In our benchmarks, Qwen3-Coder-Next_t1p95k40m0 scores 89.4/100 for agentic workflows. It runs at about 81.5 tok/s, so if you want more speed, Qwen3.8-Flash-Next-oQ3.5e-mtp is faster (~98.5 tok/s) and still scores well for agentic workflows (92.5/100).
- How fast is Qwen3-Coder-Next_t1p95k40m0 for local inference?
- Across 2 community benchmark runs, Qwen3-Coder-Next_t1p95k40m0 reaches up to 81.9 tok/s and averages 81.5 tok/s, with the fastest results on Tesla V100-PCIE-32GB.
- How much memory does Qwen3-Coder-Next_t1p95k40m0 need?
- The leanest observed configuration used about 56.8 GB of memory.
- Which tools have been used to run Qwen3-Coder-Next_t1p95k40m0?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.