LLM EXPLORER 59,358 MODELS INDEXED

Lightning 4B by quelmap

By quelmap · 13 downloads
HF Repo ↗ Copy URL

Lightning 4B is an open-source language model by quelmap. Features: 4b LLM, VRAM: 8.1GB, Context: 256K, License: apache-2.0, LLM Explorer Score: 0.33, ELO: 1328.

Base model:finetune:qwen/qwen3... Base model:qwen/qwen3-4b-think...   Conversational   En   Endpoints compatible   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/quelmap/Lightning-4b 

Lightning 4B Benchmarks

LMSys ELO 1328 vs 1272 (so35) 4.4%
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

Lightning 4B Parameters and Internals

LLM NameLightning 4B
Repository 🤗https://huggingface.co/quelmap/Lightning-4b 
Base Model(s)  Qwen3 4B Thinking 2507   Qwen/Qwen3-4B-Thinking-2507
Model Size4b
Required VRAM8.1 GB
Updated2025-09-17
Maintainerquelmap
Model Typeqwen3
Model Files  5.0 GB: 1-of-2   3.1 GB: 2-of-2
Supported Languagesen
Model ArchitectureQwen3ForCausalLM
Licenseapache-2.0
Context Length262144
Model Max Length262144
Transformers Version4.55.0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Lightning 4B

Best Alternatives
Context / RAM
Downloads
Likes
FastContext 1.0 4B SFT256K / 8.1 GB5735357
GRPO 4 70256K / 8.1 GB110
Qwen3 4B Instruct 2507256K / 8.1 GB3109972911
Fable Traces256K / 8.1 GB343210
Qwen3 4B Sft Prm256K / 17.6 GB6020
FastContext 1.0 4B RL256K / 8.1 GB455961
Qwen3 4B Thinking 2507256K / 8.1 GB291023606
...fetensors.chatgpt GPT Codex V2256K / 8.1 GB6662
...rs.chatgpt GPT Gpt5.1 Thinking256K / 8.1 GB4544
Qwen3 4B Instruct 2507 FP8256K / 5.2 GB110074478
Note: green Score (e.g. "73.2") means that the model is better than quelmap/Lightning-4b.