lexi-coder-v4.4 is an open-weight model from Reallexi LLC. It has 3.8B parameters and a 131,072-token context. At 16-bit it needs about 9.2 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index. It draws 1.2k downloads a month.
Produced by Reallexi LLC on Reallexi AI Model Builder, a local-first training platform (https://llm.reallexi.io). Hugging Face repository: reallexi/lexi-coder-v4.4. Copyright (c) 2026 Reallexi LLC. All rights reserved.
Runs On
What it takes to serve lexi-coder-v4.4 (3.8B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 7.7 GB | 9.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 3.8 GB | 4.6 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 1.9 GB | 2.3 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.
lexi-coder-v4.4 on every accelerator the SAVRN Index prices, at every precision
Model Card
Produced by Reallexi LLC on Reallexi AI Model Builder, a local-first training platform (https://llm.reallexi.io). Hugging Face repository: reallexi/lexi-coder-v4.4. Copyright (c) 2026 Reallexi LLC. All rights reserved.
Excerpt from the card by Reallexi LLC.
Configuration
- Architecture
- Phi3ForCausalLM
- Context length (tokens)
- 131,072
- Layers
- 32
- Hidden size
- 3,072
- Feed-forward size
- 8,192
- Attention heads
- 24
- Key/value heads
- 8
- Vocabulary size
- 200,064
- Sliding window (tokens)
- 262,144
- Model type
- phi3
Identity and Version
- Repository
- reallexi/lexi-coder-v4.4
- Publisher
- Reallexi LLC
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 3.8B parameters
- Languages
- Not stated by the source
- Revision
- 5ef5f75129bc9dec99728d54f8b7172cba6843fd
- First published
- 2026-09-06
- Last updated
- 2026-09-18
Files and Weights
22 files, 10.2 GB in total. The weights are 5 files totalling 10.2 GB in gguf, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| lexi-coder-v4-4-q4_k_m.gguf | Weights | 2.5 GB | 56086ac86af3 |
| model-00001-of-00004.safetensors | Weights | 2.0 GB | be21677d3027 |
| model-00002-of-00004.safetensors | Weights | 1.9 GB | fd73c1dd4e89 |
| model-00003-of-00004.safetensors | Weights | 2.0 GB | 5919bf46fe4a |
| model-00004-of-00004.safetensors | Weights | 1.8 GB | 491cb7d2321e |
| added_tokens.json | Configuration | 249 B | — |
| config.json | Configuration | 2.6 KB | — |
| generation_config.json | Configuration | 168 B | — |
| model.safetensors.index.json | Configuration | 16.3 KB | — |
| reallexi-model.json | Configuration | 626 B | — |
| samples.json | Configuration | 2.6 KB | — |
| special_tokens_map.json | Configuration | 587 B | — |
| NOTICE | Documentation | 247 B | — |
| README.md | Documentation | 364 B | — |
| SAMPLES.md | Documentation | 2.0 KB | — |
| chat_template.jinja | Other | 423 B | — |
| training_curve.png | Other | 72.0 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 2.4 MB | — |
| tokenizer.json | Tokenizer | 15.5 MB | 5ae4719cd8d4 |
| tokenizer_config.json | Tokenizer | 508 B | — |
| vocab.json | Tokenizer | 3.9 MB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 10.2 GB
Released by Reallexi LLC through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 10.2 GB |
| 16-bit | 7.7 GB |
| 8-bit | 3.8 GB |
| 4-bit | 1.9 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About lexi-coder-v4.4
How much GPU memory does lexi-coder-v4.4 need?
About 9.2 GB at 16-bit and 2.3 GB at 4-bit: the weights (3.8B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run lexi-coder-v4.4 on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
What is lexi-coder-v4.4's context length?
131,072 tokens, from the maximum position embeddings in its published configuration.