Model benchmarks
Thinklite-27-GGUF_t1p95k20m0 local LLM performance
As of November 2026, Thinklite-27-GGUF_t1p95k20m0 runs at up to 49.0 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
49.0 tok/s
Average speed
48.6 tok/s
Avg PP
723.6 tok/s
Min memory
40.9 GB
Max context
32,768 tokens
Avg output / run
25,608 tokens
Avg runtime / run
8m 56s
Avg quality
79.9
Benchmark runs
3
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 3 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 | 71.4 | 79.9 | 87.8 |
| Agent Workflow | 83.4 | 87.9 | 90.9 |
| Code Generation | 26.2 | 60.2 | 85.1 |
| Role Play & Narrative | 82.6 | 89.6 | 93.8 |
| Research & Analysis | 77.6 | 81.8 | 84.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination Thinklite-27-GGUF_t1p95k20m0 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 | — | 49.0 tok/s | 48.6 tok/s | 40.9 GB | 32,768 tokens | 79.9 | 3 |
Benchmark runs
All 3 Thinklite-27-GGUF_t1p95k20m0 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Thinklite-27-GGUF_t1p95k20m0 good for coding?
- In our benchmarks, Thinklite-27-GGUF_t1p95k20m0 scores 60.2/100 for coding. It runs at about 48.6 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 Thinklite-27-GGUF_t1p95k20m0 good for agentic (tool-using) tasks?
- In our benchmarks, Thinklite-27-GGUF_t1p95k20m0 scores 87.9/100 for agentic workflows. It runs at about 48.6 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 Thinklite-27-GGUF_t1p95k20m0 for local inference?
- Across 3 community benchmark runs, Thinklite-27-GGUF_t1p95k20m0 reaches up to 49.0 tok/s and averages 48.6 tok/s, with the fastest results on Tesla V100-PCIE-32GB.
- How much memory does Thinklite-27-GGUF_t1p95k20m0 need?
- The leanest observed configuration used about 40.9 GB of memory.
- Which tools have been used to run Thinklite-27-GGUF_t1p95k20m0?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.