whatmodel.sh / EXAONE 3.5 2.4B Instruct
EXAONE 3.5 2.4B Instruct
↗ share
+ contribute
EXAONE 3.5 2.4B Instruct is an open-weights language model from LG AI Research, released December 2024. Its published size is 2.41B parameters and a 33k-token context window. Licence: exaone.
Catalogue identity notes
Maker repo LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct @ccce25b: licence tag other (exaone). 2,405,327,360 parameters from the Hugging Face safetensors metadata. Dense, so active = total. Released 2024-12: month the maker's Hugging Face repo was created (createdAt 2024-12-01). Ollama tags exaone3.5:2.4b-instruct-q4_K_M, exaone3.5:2.4b-instruct-q8_0 are the library's instruct/chat builds of this checkpoint (tag names on the pinned tags page); download sizes are estimated from the parameter count.
$ how it compares
The LLMs ranked either side of EXAONE 3.5 2.4B Instruct, out of 828 with a score, then more from LG AI Research.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 679 |  Vicuna 13B v1.5 | 36±2 | local | ||
| 680 |  Granite 3.1 2B Instruct | 36±1 | local | ||
| 681 | EXAONE 3.5 2.4B Instruct | 36±2 | local | ||
| 682 |  LLaMA 33B | 36±1 | comparison only | ||
| 683 | Code Llama 7B Instruct | 36±1 | local | ||
| more from LG AI Research | |||||
| 127 | K-EXAONE 2.0 (750B-A37B) | 67±1 | local | ||
| 162 | K-EXAONE (236B-A23B) | 65±1 | local | ||
| 164 | EXAONE 4.5 33B | 65±2 | local | ||
$ what it's good at
coding▁▁▁▃▆██▇▆▄▃▁ #395 of 458 · 20% of its score · 4 tests
Beats DeepSeek R1 Distill Qwen 7B, Qwen3.5 2B, DeepSeek R1 Distill Llama 8B here, though they rank higher overall. Trails Qwen2.5 1.5B Instruct, WizardLM 13B V1.2 here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| CodeReviewQA: solution identification (hard), invariant accuracy | 26.7% ✎ | ▃··▁▂▂▃▂█▃▄▂ beats 17/47 | 7.7% |
| CodeReviewQA: automated code refinement, exact match rate | 3.8% ✎ | ▁▃▁▁▁▁▂▂▃█▅▄ beats 2/47 | 6.7% |
| CodeReviewQA: change localisation (hard), invariant accuracy | 0.8% ✎ | █▂▁▁▂▂▂▃▅▅▅▇ beats 8/47 | 5.2% |
| CodeReviewQA: change type recognition, invariant accuracy | 65.3% ✎ | ▁▁▁······▁▃█ beats 13/47 | 0.2% |
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #271 of 393 * · 41% of its score · 2 tests
Beats Ministral 3 14B Reasoning 2512, Ministral 3 14B, Gemma 4 E4B here, though they rank higher overall. Trails Flan-T5 XXL here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Arena-Hard (version/judge unspecified) lab | 48.2% ✎ | ▃▃▂▅▃▆▅█▅▅▂▃ beats 8/30 | 23% |
| IFEval (aggregation unspecified) lab | 73.6% ✎ | ▄·▄·▄·▄▄···█ beats 1/6 | 18% |
other tests1 not in a task · 36% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| CPTUB average (SpeakLeash, 0-5 judge scale) | 33.4% ✎ | ▁▃▂▂▃█▃▆▅▄▂▃ beats 3/95 | 36% |
$ where to run it
llama.cpp · recommended
version
- Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.macOS / Linux
curl -LsSf https://llama.app/install.sh | shWindowsirm https://llama.app/install.ps1 | iex - Run EXAONE 3.5 2.4B Instruct · needs ~1.6 GB · from huggingface.co/lmstudio-communitychat in the terminal
llama cli -hf lmstudio-community/EXAONE-3.5-2.4B-Instruct-GGUF:Q4_K_M - Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf lmstudio-community/EXAONE-3.5-2.4B-Instruct-GGUF:Q4_K_M
LM Studio
version
- Install LM Studio · Open the app once before using the lms command.
- Run EXAONE 3.5 2.4B Instruct · needs ~1.6 GB · from huggingface.co/lmstudio-community
- Or download it in the terminal
lms get https://huggingface.co/lmstudio-community/EXAONE-3.5-2.4B-Instruct-GGUF@Q4_K_M -y - Then run lms ls to see its name, and lms chat followed by that name
Ollama
version
- Install Ollama
download it (macOS / Windows) ↗
Linuxcurl -fsSL https://ollama.com/install.sh | sh - Run EXAONE 3.5 2.4B Instruct · needs ~1.95 GB · from ollama.comchat in the terminal
ollama run exaone3.5:2.4b-instruct-q4_K_M
MLX (Mac)
version
- Install MLX (Mac) · Apple silicon Macs only, text only.with uv
uv tool install mlx-lmorpip install mlx-lm - Run EXAONE 3.5 2.4B Instruct · needs ~1.4 GB, Mac only · from huggingface.co/alexgusevskichat in the terminal
mlx_lm.chat --model alexgusevski/EXAONE-3.5-2.4B-Instruct-q4-mlx - Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model alexgusevski/EXAONE-3.5-2.4B-Instruct-q4-mlx
Weights and documentation: huggingface.co/LGAI-EXAONE ↗
$ how the score was worked out · LG AI Research vendor-reporte… 41%, SpeakLeash CPTUB (Complex Pol… 36%, CodeReviewQA: The Code Review… 20%
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards) · lab-reported | 2 | 41% |
| SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark) | 1 | 36% |
 CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2) | 4 | 20% |
| prior · pull toward the middle | 2% | |
| weighted average | = 36.2 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json