# H2O-Lightning-4B model sizes and sources

> H2O-Lightning-4B is an open decision model built on Qwen3.5-4B that runs on unmodified vLLM. Version 1.2.2 takes up to four images per request.

Updated 2026-11-04 (Australia/Adelaide)

By H2O.ai · Released: 2026-10-07 · OPEN

A 4B open model built on Qwen3.5-4B that runs on unmodified vLLM. Version 1.2.2 takes up to four images per request.

## Facts

- Size: 4B
- Licence: Apache-2.0

## Claims and sources

- **VENDOR-REPORTED**: Its text-only v1.1 ranked #1 of 104 open-weight systems on [JevBench](<https://decisionmodels.ai/jevbench>) v1.6.1's composite, 72.5 to [Jev](<https://decisionmodels.ai/models/jev>) 1.13's 71.5, read 7 Oct 2026 [source](<https://h2o.ai/blog/2026/h2o-lightning-jevbench>)

## Links

- [Source: Announcement](<https://h2o.ai/blog/2026/h2o-lightning-jevbench>)
- [Source: Hugging Face](<https://huggingface.co/h2oai/h2o-lightning-4b>)

## Related projects

### Decision model kernels: a faster llama.cpp on one GPU

[Decision model kernels: a faster llama.cpp on one GPU](<https://decisionmodels.ai/projects#project-decision-model-kernels>)

[Source: WyrmKeep/decision-model-kernels](<https://github.com/WyrmKeep/decision-model-kernels>)

A lab notebook for speeding up one decision model, H2O Lightning 4B, served through llama.cpp on an RTX 3060 Ti. Each request takes 2.13 to 2.45 times its roofline floor. The first measured change, a llama.cpp pull request, cut latency by 1.94% with bit-for-bit identical answers; the bigger gains in its list of 15 targets are labelled predictions.

- Creator: WyrmKeep

- Stars: 0

- Licence: MIT

- Model: [H2O-Lightning-4B](<https://decisionmodels.ai/models/h2o-lightning>)

- Tags: inference, llama.cpp, cuda