Haste Jev 20m (Base)
Role: Largest local prototype
Open-weights System-1 decision engine for software paths that need typed decisions under a time budget (agent routing, tool routing, intent classification, pre-flight guardrails, browser action selection). Not a chat model.
Measured specification
| Field | Value |
|---|---|
| Total parameters | 20,383,267 |
| Trainable parameters | 3,606,051 |
| Hash-table buffers | 16,777,216 |
| d_model | 256 |
| Layers | 4 |
| Heads | 4 |
Parameter counts match the GitHub README table (verified with verify_claims.py).
Honest claims
| Claim | Status |
|---|---|
| PICA option-order bias = 0.0% | Verified architecturally |
| Exact parameter table | Verified |
| FP32 ~0.4 MB for 100k weights | Verified (weight storage only) |
| p99 < 15ms / ECE < 0.009 / 99.4% arithmetic | Not verified — do not cite from this card |
| Published latency / accuracy on your workload | Measure yourself |
Weights may be lightly or untrained prototypes depending on export; treat behavioral accuracy as unknown until you evaluate on labeled data.
Quickstart
from hastejev import HasteJevEngine
eng = HasteJevEngine.from_pretrained("noffy/hastejev")
# eng = HasteJevEngine.from_pretrained("noffy/hastejev", quantization="int4")
r = eng.choice(
"Request: reset password for user@corp.example",
["auth_self_service", "billing", "security_review"],
)
print(r.decision, r.confidence)
Install: pip install git+https://github.com/racstan/hastejev.git
Files
| File | Contents |
|---|---|
model.safetensors / pytorch_model.bin | FP32 state dict |
model_fp16.safetensors | FP16 |
model_int8.safetensors | True weight-only int8 (weight_q) when re-exported with ≥1.1.0 |
model_int4.safetensors | True packed int4 (weight_packed) when re-exported with ≥1.1.0 |
Older revisions of model_int8/model_int4 may be mislabeled FP32; re-export or
re-download after this commit.
License
Apache-2.0