Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash
Cactus Needle 3 introduces automation models ranging from 8-29MB, capable of matching DeepSeek V4 Flash. These models feature an "intelligence ladder" design, where a single set of weights supports various depths from 2 to 20 layers. Subnetworks as small as 2 layers can be fine-tuned for specific tasks, achieving frontier-level accuracy on devices smaller than the full model. Fine-tuning on DroidCall, for instance, improves every subnetwork by 18 to 36 points, with 4-layer subnetworks and above surpassing DeepSeek V4 Flash, starting at 29M parameters.
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
IngestedOffset at this time: UTC+0Sep 18, 2026, 17:00 UTC
- Ingested
- Sep 18, 2026, 17:00
- Source type
- Unclassified
Discussion trend
The percentage is based on collected discussion signal, not new comments or independent people. The curve only compares the same topic across time.
- Basis
- Running about 2.4× the median of this source's recent listed items
- Metric comparison
- 120 vs median 50.5 (20 baseline samples)
- Detected
- 09/18, 22:00
Full text isn't available here.
Read at source →