Model benchmarks
flint local LLM performance
As of October 2026, flint runs at up to 195.9 tok/s for local inference (best of 4 community benchmark runs across 3 GPUs).
Model size
40B
Peak speed
195.9 tok/s
Average speed
118.9 tok/s
Avg PP
1472.3 tok/s
Min memory
29.7 GB
Max context
262,144 tokens
Avg output / run
23,507 tokens
Avg runtime / run
7m 24s
Avg quality
71.9
Benchmark runs
4
GPUs tested
3
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 4 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 | 66.0 | 71.9 | 76.7 |
| Agent Workflow | 76.5 | 78.4 | 82.3 |
| Code Generation | 29.5 | 53.3 | 68.5 |
| Role Play & Narrative | 80.8 | 85.6 | 87.6 |
| Research & Analysis | 62.6 | 70.4 | 78.9 |
Performance by hardware and tool
Every hardware/tool/quantization combination flint 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 AI PRO R9700 | Strata | — | 195.9 tok/s | 191.3 tok/s | n/a | 262,144 tokens | 75.4 | 2 |
| NVIDIA GeForce RTX 3080 | Strata | — | 70.2 tok/s | 70.2 tok/s | 29.7 GB | 8,192 tokens | 65.0 | 1 |
| AMD Radeon RX 7600 XT | LM Studio | Q4_K_M | 22.6 tok/s | 22.6 tok/s | n/a | 65,536 tokens | 71.9 | 1 |
Benchmark runs
All 4 flint runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is flint good for coding?
- In our benchmarks, flint scores 53.3/100 for coding. It runs at about 118.9 tok/s, so if you want more speed, Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well for coding (82.2/100).
- Is flint good for agentic (tool-using) tasks?
- In our benchmarks, flint scores 78.4/100 for agentic workflows. It runs at about 118.9 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~162.4 tok/s) and still scores well for agentic workflows (87.8/100).
- How fast is flint for local inference?
- Across 4 community benchmark runs, flint reaches up to 195.9 tok/s and averages 118.9 tok/s, with the fastest results on AMD Radeon AI PRO R9700.
- How much memory does flint need?
- The leanest observed configuration used about 29.7 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run flint?
- Benchmarks were submitted using LM Studio, Strata. Results are community-contributed and updated as new runs arrive.