Decision Workspace
tq-kv vs turboquant-rs vs qatq
Side-by-side comparison of Rust crates
50
tq-kv
experimentalv0.6.0
TurboQuant KV Cache Compression for LLMs — First working on GGUF Q4_K_M (3-Fix framework). KV Compaction (50x token reduction). Per-head adaptive bitwidth. 4-bit value compression. Calibration pipeline. SRHT QJL (115x faster). AVX2 SIMD fused attention. C FFI. 114 tests.
51
turboquant-rs
experimentalv0.4.1
TurboQuant KV-Cache Quantization — 3-bit compression with zero accuracy loss (Zandieh et al., ICLR 2026)
49
qatq
experimentalv0.4.2
Exact tensor-aware compression for LLM KV caches and runtime migration
Core Metrics
| tq-kv | turboquant-rs | qatq | |
|---|---|---|---|
| Health Score | 50 | 51 | 49 |
| Total Downloads | 264 | 5.4K | 1.3K |
| 30d Downloads | 0 | 0 | 0 |
| Dependents | 0 | 5 | 5 |
| Releases | 6 | 5 | 12 |
| Last Updated | 192d ago | 173d ago | 69d ago |
| Age | 6m | 6m | 3m |
Health Breakdown
tq-kv
Maintenance
15
Quality
16
Community
6
Popularity
3
Documentation
10
turboquant-rs
Maintenance
11
Quality
13
Community
8
Popularity
4
Documentation
15
qatq
Maintenance
15
Quality
10
Community
8
Popularity
4
Documentation
12
Technical Details
| tq-kv | turboquant-rs | qatq | |
|---|---|---|---|
| Version | 0.6.0 | 0.4.1 | 0.4.2 |
| Stable (≥1.0) | ✗ No | ✗ No | ✗ No |
| License | MIT OR Apache-2.0 | MIT OR Apache-2.0 | Apache-2.0 |
| Dependencies | 6 | 9 | 7 |
| Crate Size | 76KB | 144KB | 1.5MB |
| Features | 5 | 4 | 3 |
| Yanked % | 0.0% | 0.0% | 8.3% |
| Edition | 2021 | 2021 | 2024 |
| MSRV | — | 1.75 | 1.96 |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •turboquant-rs leads with a health score of 51/100, but none of the options score above 80.