Sponsored by
Vmake Labs Video Enhancer
-Vmake Labs Video Enhancer uses AI to improve video quality, restore details, reduce blur, and upscale videos to 8K.
INT4 (GPTQ) build of
Eikos-27B
: 4-bit weights (GPTQ W4A16, groups of 128) with bf16 activations, made with llm-compressor (compressed-tensors format), calibrated on 256 training items (never on evaluation data).
The vision tower, MTP weights, embeddings and LM head are kept in higher precision. The prompt format, the
letter readout and the calibration (
calib.json
, T = 1) are the same as the bf16 model.
Eikos answers typed decisions (yes/no, one of N options, ordinal scores) about a given state in one forward pass,
with a calibrated probability for every option. See the
Eikos-27B card
for what the
model does, how it was trained, the full evaluation and its limitations.
Use
Requires vLLM ≥ 0.30.0.
Older builds return wrong answers when several long requests are batched together on
this hybrid (Gated DeltaNet) architecture.
The HTTP API, agent sessions and the question types are the same as for
Eikos-27B
.
Validation against bf16
Same 7,371 items for both builds (7 suites, never used in training), vLLM 0.30 with batching and prefix cache on.
The release gate was fixed before looking at results: accuracy within 1 point of bf16, ECE within 0.01, and at
least 97% of answers unchanged.
Eikos-27B (bf16)
Eikos-27B-INT4
Size
55.6 GB
19.4 GB
JevBench public — original / hard
100.0 / 82.0
100.0 / 82.9
DecisionBench — medium / hard
89.1 / 78.2
88.4 / 78.5
General battery (9 tasks)
82.6
82.5
Finance (CUAD, sentiment, FinQA-judge)
85.4
84.9
Trade rules — seen / unseen
85.4 / 87.4
85.4 / 88.6
Compositional rules — same type / new domain / rulebooks
95.6 / 94.3 / 95.3
95.6 / 94.5 / 95.3
ECE (lower is better)
0.043
0.041
≥0.90 confidence: decides / error
44.1% / 2.5%
43.6% / 2.4%
Same answer as bf16 (all / confident ≥0.9)
—
97.8% / 100.0%
It passes our release gate.
License
MIT for our contributions (
LICENSE
). The base model, Qwen3.8-27B, is Apache-2.0 (
LICENSE-Qwen
);
attributions are in
NOTICE
. Not legal, tax or investment advice.
Runs of caiovicentino1 Eikos-27B-INT4 on huggingface.co
315
Total runs
1
24-hour runs
7
3-day runs
31
7-day runs
294
30-day runs
More Information About Eikos-27B-INT4 huggingface.co Model
Eikos-27B-INT4 huggingface.co is an AI model on huggingface.co that provides Eikos-27B-INT4's model effect (), which can be used instantly with this caiovicentino1 Eikos-27B-INT4 model. huggingface.co supports a free trial of the Eikos-27B-INT4 model, and also provides paid use of the Eikos-27B-INT4. Support call Eikos-27B-INT4 model through api, including Node.js, Python, http.
Eikos-27B-INT4 huggingface.co is an online trial and call api platform, which integrates Eikos-27B-INT4's modeling effects, including api services, and provides a free online trial of Eikos-27B-INT4, you can try Eikos-27B-INT4 online for free by clicking the link below.
caiovicentino1 Eikos-27B-INT4 online free url in huggingface.co:
Eikos-27B-INT4 is an open source model from GitHub that offers a free installation service, and any user can find Eikos-27B-INT4 on GitHub to install. At the same time, huggingface.co provides the effect of Eikos-27B-INT4 install, users can directly use Eikos-27B-INT4 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.