awesome-latest-LLM

最新LLMの一覧を作成します
https://github.com/stardust-coder/awesome-latest-LLM

Last synced: 3 days ago
JSON representation

Dataset
- Corpus
  - Healthcare Datasets (Aloe Beta)
  - PMC-Patients
- Evaluation benchmarks
- Image + Text
  - VQA-RAD
  - Clinical NLP 2023
  - MedICaT
  - MedVTE
  - MedAlign(Stanford)
  - ECG-QA
  - He et al.(2023)
  - MedTrinity
  - MedEval
  - MedLLMsPracticalGuide
  - Medical datasets for LLMs (collection)
  - MIMIC-ECG-IV - caption dataset
  - OmniMedVQA - items, covering 12 different medical image modalities and referring to more than 20 human anatomical regions.
  - LLaVA-Med Dataset - 4 to generate diverse biomedical multimodal instruction-following data using image-text pairs from PMC-15M.
  - PMC-OA - caption pairs
  - SLAKE - answer pairs
  - PathVQA
  - Awesome-Medical-Dataset
- Only Text
  - MedMCQA
  - PubMedQA
  - MedQA (USMLE)
  - PubHealth
  - MMLU
  - K-Q&A
  - HeadQA
  - LongHealth
  - Medical Eval Sphere
  - MedCalcBench - Bench-v1.0)
  - PMC Patients
  - MedQA-Calc
  - MedS-Bench
  - MedQuAD
  - TJH Dataset
  - **FreedomIntelligence**
  - Medical O1 Reasoning
  - Medical O1 Verifiable Problem
  - Disease Database
  - **OnDeviceMedNotes**
  - synthetic-medical-conversations-deepseek-v3
  - EquityMedQA - ended Q&A for equity and bias mitigation.
  - MedDistractQA
  - HealthsearchQA
  - MeDiSumQA - IV. released on Physionet.
  - MedNLI - III dataset, logical relationship between a premise and a hypothesis
  - MeQSum
  - MIMIC-IV
  - ClinicBench
  - AlpaCare-MedInstruct-52k
  - LiveQA
  - HealthBench (OpenAI)
  - CLUE
  - CUREBench (Harvard)
  - GlobMed
  - OpenMed
  - MultiMedX
- - He et al.(2023)
  - J-ResearchCorpus
  - Apollo Corpus JP
  - MIMIC-ECG-IV - caption dataset
  - JMMLU - translated version of MMLU
  - IgakuQA（Japanese National Medical License Exam）
  - JMedData4LLM
English-centric
- Amber - 2.0 | Llama|| totally open|
- Phi-1.5(Microsoft) - 1_5) | 1.3B| MSRA-license||textbooks|
- Reka Flash
- Miqu - 1-70b/tree/main) | 70B | none ||| leaked from Mistral |
- Gemma(Google)
- Aya(Cohere) - 101) | 13B | apache-2.0 | || multilingual |
- Command-R(Cohere) - command-r-v01) | 35B | non commercial | || RAG capability |
- BTX(Meta)
- Mixtral-8x22B(Mistral) - community/Mixtral-8x22B-v0.1) | 8x22B | apache-2.0 | || MoE |
- Phi-3(Microsoft) - 3-medium-128k-instruct) | 3.8B, 13B | MIT | Phi-3 datasets | - | |
- Mixtral-8x7B - 8x7B-Instruct-v0.1) | 8x7B | apache-2.0 |||MoE, [offloading](https://github.com/dvmazur/mixtral-offloading)|
- Grok-1
- LongNet(Microsoft) - | apache-2.0 | [MAGNETO](https://arxiv.org/pdf/2210.06423.pdf)| input 1B token| |
- gigaGPT(Cerebras) - 2.0 | | |
- Mamba - spaces/mamba-2.8b) | 2.8B | apache-2.0 | based on state space model| |
- QWen(Alibaba) - 72B) | 72B | [license](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT)| 3T tokens | | beats Llama2 |
- Self-RAG - 2.0 | 13B | | | critic model |
- TinyLlama - 1.1B-intermediate-step-1431k-3T) | apache-2.0 | 1.1B | based on Llama, 3T token | | |
- Xwin-LM - LM/Xwin-LM-70B-V0.1) | 70B | Llama2 |based on Llama2| also codes and math|
- BTX(Meta)
- DeepSeek-V3 - ai/DeepSeek-V3) | 671B | [link](https://github.com/deepseek-ai/DeepSeek-V3/blob/main/LICENSE-MODEL) | 14.8T | sft, RL | MoE |
- Phi-4 (Microsoft) - 4) | 14B | msrla | | | small, sft, dpo |
- Minimax-01 - Text-01) | [Minimax](https://github.com/MiniMax-AI/MiniMax-01?tab=License-1-ov-file) | 456(45.9)B | 1M token context length | | MoE, 4M token window |
- DeepSeek-R1 - ai/DeepSeek-R1)| 671B | MIT | | |
- Llama4 (Meta) - llama/llama-4-67f0c30d9fe03840bc9d0164)|17B|llama4|30T token||10M token|
- Awesome-LLM
- Llama2(Meta) - llama) | 70B | Llama2 | 2T tokens| chat-hf seems the best|
- Olmo 3 (Allen) - 3) | 7, 32B | apache-2.0 | | | |
- OpenVLM Leaderboard
- rnj-1(EssentialAI) - 1-instruct) | 8B | apache2.0 | 8.4T+380B tokens | 150B tokens | code and STEM |
- Mistral-Large-3 - Large-3-675B-Instruct-2512) | 675B | | | | |
- Kimi-K2.5 - K2.5) | 1TA32B | modifiedMIT | Kimi-K2-Base | 15T tokens | moe |
- GLM 5.1 - org/GLM-5.1) | 754B | MIT | | | |
- Llama 3(Meta) - llama/Meta-Llama-3-70B-Instruct) | 70B | [META LLAMA3](https://llama.meta.com/llama3/license/) | || [extended to 120B](https://huggingface.co/mlabonne/Meta-Llama-3-120B-Instruct) |
- Mixtral-8x22B(Mistral) - community/Mixtral-8x22B-v0.1) | 8x22B | apache-2.0 | || MoE |
- Kimi K3 - K3) | 2.8TB | [Kimi K3 License](https://huggingface.co/moonshotai/Kimi-K3/blob/main/LICENSE) | ? | ? | 1T context window, moe(104Ba) |
- FIM-Midtraining (TIGER AI Lab) - Lab/fim-midtraining) | 14B | Apache-2.0 | Qwen2.5-Coder / Qwen3 with function-aware fill-in-the-middle mid-training | R2E-Gym / SWE-Smith / SWE-Lego | [paper](https://arxiv.org/abs/2607.12463)|
- Phi-4 (Microsoft) - 4) | 14B | msrla | | | small, sft, dpo |
- Miqu - 1-70b/tree/main) | 70B | none ||| leaked from Mistral |
- Amber - 2.0 | Llama|| totally open|
Evaluation
- 医療ドメイン特化LLMの性能はどうやって評価する？
- Japanese Medical Language Model Evaluation Harness
Japanese-centric
- ELYZA-japanese-Llama-2-13b - japanese-Llama-2-13b) | 13B | | Llama-2-13b-chatベース |
- Swallow(東工大) - llm) | 70B | | Llama2-70Bベース |
- StableLM(StabilityAI) - stablelm-base-beta-70b) | 70B | | Llama2-70Bベース |
- KARAKURI 70B - ai/karakuri-lm-70b-v0.1) | 70B | cc-by-sa-4.0 | Llama2-70Bベース | | [note](https://note.com/ngc_shj/n/n46ced665b378?sub_rt=share_h)|
- LLM-jp - jp) | 13B | DPO追加あり |
- LLama3ELYZA-JP-8B - 3-ELYZA-JP-8B) | 8B | Llama3 | Llama3 | | 70B not open |
- KARAKURI LM 8x7B - ai/karakuri-lm-8x7b-chat-v0.1) | 8x7B | Apache-2.0 | | | MoE |
- awesome-japanese-llm - jp.github.io/awesome-japanese-llm/) and [日本語LLM評価](https://swallow-llm.github.io/evaluation/about.ja.html)
- PlaMo 2 (PFN) - 2-8b) | 8B | [plamo](https://tech.preferred.jp/ja/blog/plamo-community-license/) ||| Samba |
- PLaMo 2.0-31B
- KARAKURI 70B - ai/karakuri-lm-70b-v0.1) | 70B | cc-by-sa-4.0 | Llama2-70Bベース | | [note](https://note.com/ngc_shj/n/n46ced665b378?sub_rt=share_h)|
- KARAKURI LM 8x7B - ai/karakuri-lm-8x7b-chat-v0.1) | 8x7B | Apache-2.0 | | | MoE |
Leaderboard
Model
- AMIE(Google) - | - | based on PaLM 2 | | | EHR|
- Med-PaLM M(Google) - | PaLM2 | | |multi-modal|
- Med-PaLM2(Google) - | PaLM2 | | |
- Med-PaLM(Google) - | PaLM | | | |
- BioMistral - | | | | |
- Health-LLM(Rutgersなど)
- JMedLoRA(UTokyo) - CVM-utokyohospital/llama2-jmedlora-3000) | 70B | none | none | QLoRA | IgakuQA | Japanese, insufficient quality |
- Almanac(Stanford) - davinci-003 | | | RAG |
- AdaptLLM(Microsoft Research) - LLM) | 7B, 13B | | reading comprehensive corpora | | | | ICLR2024 |
- Hippocrates
- AdaptLLM(Microsoft Research) - LLM-13B) | 7B, 13B | | reading comprehensive corpora | | | | ICLR2024 |
- Med-Gemini(Google) - | Gemini | | |multimodal|
- BiMediX - commercial | 8x7B | mixtral8x7B | | | MoE |
- Meditron(EPFL) - | 8B | - | Llama3 | | MedQA, MedMCQA, PubmedQA | SOTA |
- HF - 2.0 | Llama3 | 100,000+ data, [ORPO](https://huggingface.co/blog/mlabonne/orpo-llama-3) | | |
- BioMistral - | | | | |
- Apollo - 7B) | ~7B | | | | | | multilingual |
- Meditron(EPFL) - llm/meditron-70B) | 70B | Llama2 | Llama2 | GAP-Replay(48.1B) | [dataset](img/meditron-testdata.png),[score](img/meditron-eval2.png) | |
- BioMedGPT(Luo et al.)
- PMC-LLaMa
- Med-Flamingo
- LLaVa-Med(Microsoft) - med-7b-delta) | 13B | - | LLaVa| medical dataset | VAQ-RAD, SLAKE, PathVQA |multi-modal|
- Awesome-Healthcare-Foundation-Models
- UltraMedical(TsinghuaC3I) - | Llama3 | | | |
- MedLLMsPracticalGuide
- 医療分野に特化したLLM紹介
- Huatuo-o1 - o1-72B) | 72B | apache-2.0 |
- Health-LLM(Rutgersなど)
- Meditron(EPFL) - | 8B | - | Llama3 | | MedQA, MedMCQA, PubmedQA | SOTA |
- OpenMeditron - 70B) | 7~70B | |||MedQA etc. |
- Awesome-Medical-Large-Language-Models
- Awesome-Medical-LLM
- OmniV-Med(Alibaba)
- ELYZA-Med-Base-1.0-Qwen2.5-72B
- Med-R1 (IEEE) - R1)| 2B | | Qwen2-VL | | | VLM |
- Med-R1 8B (IQVIA)
- JPharmatron(EQUES) - 7B) | 7B | cc-by-sa-4.0 | Qwen2.5 | pharma corpus | None | Japanese, AACL2025 |
- AMIE(Google) - | - | based on PaLM 2 | | | EHR|
- Almanac(Stanford) - davinci-003 | | | RAG |
- ChatGPT Health (OpenAI)
- MedGemma (Google) - release-680aade845f90bec6a3f60c4)| 4B, 27B | | Gemma3 | | | |
- ChatGPT for Clinicians (OpenAI)
- MeditronFO (EPFL) - 2.0, dataset is NonCommercial. | Apertus、OLMo、EuroLLM | fullSFT with QA | fully open |
Small language models (SLM)
- PLaMo 2 2B - 2.0| | | pruning, tested on HumanEval+ |
- Phi-4 mini
- Sarashina2.2 - tasks=3.75 |
- SmolLM (Huggingface) - 6723884218bcda64b34d7db9)| 135M~1.7B| apache-2.0 | |
- OLMo-2 - 2-0425-1B-Instruct) | 1B | | | |
- Bonsai (PrismML) - ml/bonsai) | 1.7~8B | apache2.0| | |
- Ling-3.0-tiny (inclusionAI) - 3.0-tiny) | 7.9B | mit| ? | ? | moe 1.3Ba |
Table of Contents
- Qwen3-Omni (Alibaba) - omni-68d100a86cd0906843ceccbe) | 30B-A3B | apache-2.0 | text-first pretraining and mixed multimodal training | | [demo](https://huggingface.co/spaces/Qwen/Qwen3-Omni-Demo) |
- Uni-Moe-Omni (HIT) - TMG/Uni-MoE-2.0-Omni) | 33B-1.5~18B | apache-2.0 | 75B token | | MoE, surpass Qwen2.5-Omni |
- Ministral 3
- Qwen3-Next-80B-A3B - Next-80B-A3B-Instruct-GGUF)
- Devestral 2
- Fara (Microsoft) - 7B) | 7B | mit | Qwen2.5-VL-7B | | |
- GLM-4.7
- FunctionGemma (Google) - org/AutoGLM-Phone-9B-Multilingual) | 0.27B | | | | function calling |
- AutoGLM-Phone-9B-Multilingual (ZAI) - org/AutoGLM-Phone-9B-Multilingual) | 9B | mit (for research and educational purposes only.) | | | smartphone |
Uncategorized
- Uncategorized

Programming Languages

Python 31 HTML 1 TypeScript 1

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Awesome

awesome-latest-LLM

Dataset

Corpus

Evaluation benchmarks

Image + Text

Only Text

English-centric

Evaluation

Japanese-centric

Leaderboard

Model

Small language models (SLM)

Table of Contents

Uncategorized

Uncategorized