Model benchmarks
spark-2.6-4b local LLM performance
As of November 2026, spark-2.6-4b runs at up to 144.7 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
4B
Peak speed
144.7 tok/s
Average speed
144.3 tok/s
Avg PP
4216.5 tok/s
Min memory
4.2 GB
Max context
32,768 tokens
Avg output / run
6,750 tokens
Avg runtime / run
59s
Avg quality
54.0
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 | 50.3 | 54.0 | 57.6 |
| Agent Workflow | 75.6 | 77.9 | 80.1 |
| Code Generation | 38.4 | 42.3 | 46.1 |
| Role Play & Narrative | 57.0 | 59.7 | 62.4 |
| Research & Analysis | 25.7 | 36.1 | 46.4 |
Performance by hardware and tool
Every hardware/tool/quantization combination spark-2.6-4b 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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 3080 Ti | OpenAI-compatible | — | 144.7 tok/s | 144.3 tok/s | 4.2 GB | 32,768 tokens | 54.0 | 2 |
Benchmark runs
All 2 spark-2.6-4b runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is spark-2.6-4b good for agentic (tool-using) tasks?
- In our benchmarks, spark-2.6-4b scores 77.9/100 for agentic workflows. It runs at about 144.3 tok/s, so if you want more speed, unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M is faster (~246.8 tok/s) and still scores well for agentic workflows (80.6/100).
- How fast is spark-2.6-4b for local inference?
- Across 2 community benchmark runs, spark-2.6-4b reaches up to 144.7 tok/s and averages 144.3 tok/s, with the fastest results on NVIDIA GeForce RTX 3080 Ti.
- How much memory does spark-2.6-4b need?
- The leanest observed configuration used about 4.2 GB of memory.
- Which tools have been used to run spark-2.6-4b?
- Benchmarks were submitted using OpenAI-compatible. Results are community-contributed and updated as new runs arrive.