Decision Workspace
inferlab vs newt-inference vs vllm-cpp
Side-by-side comparison of Rust crates
48
inferlab
experimentalv0.21.1
Reproducible LLM inference experiment harness: versioned workspaces, closed-loop serve/eval/bench recipes, runtime images, and file-first evidence.
53
newt-inference
experimentalv0.7.5
Newt-Agent inference layer: backend trait + local Ollama/vLLM + provider-plugin host
49
vllm-cpp
experimentalv0.0.1
Safe model inference, streaming, chat, and concurrent requests for vllm.cpp
Core Metrics
| inferlab | newt-inference | vllm-cpp | |
|---|---|---|---|
| Health Score | 48 | 53 | 49 |
| Total Downloads | 741 | 407 | 15 |
| 30d Downloads | 0 | 0 | 0 |
| Dependents | 0 | 22 | 0 |
| Releases | 43 | 5 | 1 |
| Last Updated | 2d ago | 72d ago | 46d ago |
| Age | 3m | 4m | 1m |
Health Breakdown
inferlab
Maintenance
19
Quality
8
Community
6
Popularity
3
Documentation
12
newt-inference
Maintenance
15
Quality
14
Community
9
Popularity
3
Documentation
12
vllm-cpp
Maintenance
10
Quality
17
Community
6
Popularity
1
Documentation
15
Technical Details
| inferlab | newt-inference | vllm-cpp | |
|---|---|---|---|
| Version | 0.21.1 | 0.7.5 | 0.0.1 |
| Stable (≥1.0) | ✗ No | ✗ No | ✗ No |
| License | MIT | Apache-2.0 | MIT OR Apache-2.0 |
| Dependencies | 44 | 24 | 5 |
| Crate Size | 995KB | 81KB | 63KB |
| Features | 0 | 5 | 11 |
| Yanked % | 0.0% | 0.0% | 0.0% |
| Edition | 2024 | 2021 | 2021 |
| MSRV | 1.89 | 1.88 | 1.85 |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •newt-inference leads with a health score of 53/100, but none of the options score above 80.
- •inferlab has the most downloads (741), suggesting wider adoption.