Open weights
37 milestones sit on this thread. For 2 of them it is the main thread. The others also belong to another thread, which is named under them.
2020
- The PileThe Pile: An 800GB Dataset of Diverse TextMain thread: Language modellingDataset
2021
- T0Multitask Prompted Training Enables Zero-Shot Task GeneralizationMain thread: AlignmentModel
2022
- OPT-175BOPT: Open Pre-trained Transformer Language ModelsMain thread: Scale and dataModel
- BLOOMBLOOM: A 176B-Parameter Open-Access Multilingual Language ModelMain thread: Language modellingModel
- GPTQGPTQ: Accurate Post-Training Quantization for Generative Pre-trained TransformersMain thread: Efficiency and infrastructureTechnique
2023
- LLaMALLaMA: Open and Efficient Foundation Language ModelsMain thread: Language modellingModel
- llama.cppllama.cppMain thread: Efficiency and infrastructureInfrastructure
- AlpacaStanford AlpacaMain thread: AlignmentModel
- QLoRAQLoRA: Efficient Finetuning of Quantized LLMsMain thread: Efficiency and infrastructureTechnique
- FalconFalcon / RefinedWeb modelsMain thread: Language modellingModel
- AWQAWQ: Activation-aware Weight QuantizationMain thread: Efficiency and infrastructureTechnique
- Llama 2Llama 2: Open Foundation and Fine-Tuned Chat ModelsMain thread: Language modellingModel
- Mistral 7BMistral 7BMain thread: Language modellingModel
- Mixtral 8x7BMixtral of ExpertsMain thread: Language modellingModel
2024
- DeepSeek LLMDeepSeek LLM: Scaling Open-Source Language Models with LongtermismMain thread: Scale and dataResearch paper
- DolmaDolma: an Open Corpus of Three Trillion TokensMain thread: Scale and dataDataset
- OLMoOLMo: Accelerating the Science of Language ModelsMain thread: Language modellingModel
- GemmaGemma: Open Models Based on Gemini Research and TechnologyModel
- Llama 3Llama 3Main thread: Scale and dataModel
- Phi-3Phi-3 Technical ReportMain thread: Scale and dataModel
- DeepSeek-V2 / MLADeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language ModelMain thread: Efficiency and infrastructureModel
- FineWebThe FineWeb Datasets: Decanting the Web for the Finest Text Data at ScaleMain thread: Scale and dataDataset
- Qwen2Qwen2 Technical ReportMain thread: MultimodalityModel
- Llama 3.1 405BThe Llama 3 Herd of ModelsMain thread: Scale and dataModel
- DeepSeek-V3DeepSeek-V3 Technical ReportMain thread: Scale and dataModel
2025
- DeepSeek-R1DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement LearningMain thread: Language modellingModel
- Llama 4Llama 4Main thread: Efficiency and infrastructureModel
- Qwen3Qwen3Main thread: ReasoningModel
- Kimi K2Kimi K2Main thread: Language modellingModel
- gpt-ossgpt-oss-120b and gpt-oss-20bMain thread: ReasoningModel
- GLM-4.5GLM-4.5Main thread: ReasoningModel
- DeepSeek-V3.2DeepSeek-V3.2 with DeepSeek Sparse AttentionMain thread: Language modellingModel
2026
- Qwen3.5Qwen3.5Main thread: ReasoningModel
- DeepSeek-V4DeepSeek-V4Main thread: Efficiency and infrastructureModel
- GLM-5.2GLM-5.2Main thread: ReasoningModel
- Kimi K3Kimi K3Main thread: MultimodalityModel
- Qwen3.8Qwen3.8Model