Alternatives to Magic.dev

Open-weight frontier models trained with reinforcement learning for domain knowledge work in finance, law, consulting and healthcare.

Falcon LLM: Apache 2.0 open-weight models for Arabic, Emirati and multimodal AI you self-host

Zhipu GLM delivers open-weight GLM-5.3 models with 1M-token context, MaaS APIs, and shipped computer-use agents.

Open-weight agentic coding models: Laguna S 2.1 (118B) and on-device Laguna XS 2.1 (33B).

KoBigBird brings long-document Korean NLP up to 4096 tokens with BigBird block sparse attention.

PaLM API gives developers safe, scalable access to Google's language models, paired with MakerSuite for prompt prototyping.

StableLM: Stability AI's open-weight language model for self-hosted text and code generation

RobBERT is a family of open pre-trained Dutch language models you fine-tune for classification, regression, and token-tagging tasks.

Pay-per-token access to MiniMax M3 and M3.1 Flash language APIs plus H3 video, Speech 2.8 and Music 3.0.

Transformers is Hugging Face's open-source model-definition framework for running, fine-tuning, and serving text, vision, audio, and

Qwen3.6-27B: Apache 2.0 agentic coding and vision reasoning on your own GPU.

AI21 post-trains specialized models on your proprietary data for frontier-level quality at lower cost per call

MGM-Omni: open-weight omni-modal LLM demo fusing speech, text, and vision with zero-shot voice cloning.

Open-source add-on to Hugging Face Transformers for training, composing, and sharing parameter-efficient fine-tuning adapters.

Open-weight multimodal model built for agents that read images and then call tools on what they see.

HPT is HyperGAI's open-weight multimodal vision-language LLM for edge devices and self-hosted inference.

Open-source multilingual BERT for 11 Indic languages and Indian-English.

Browse 2M+ open models, 500k+ datasets and 1M+ Spaces on the Hugging Face machine learning hub

Mistral: open-weight AI models you can self-host, fine-tune and run in-region

Cohere delivers enterprise RAG and sovereign AI models you can run in your own VPC or on-premises.

ModelScope is Alibaba's open model hub for hosting, fine-tuning, evaluating and demoing Qwen, MiniMax and DeepSeek weights.

SmallDoge small language models: open SLM weights, training checkpoints, and SLM datasets on Hugging Face

StepFun's API delivers OpenAI-compatible frontier reasoning, vision-language document parsing and a full StepAudio 3 voice stack.

Ludwig trains multi-modal and LLM models from a single YAML config file — no training loops, Apache 2 licensed.

Feyn trains custom AI models on your proprietary data so you own the weights and keep improving in production.

NXAI builds xLSTM memory architectures for edge and industrial time-series work, shipping TiRex-1 and TiRex-2 models on Hugging Face and GitHub.

Keras-native pretrained models for BERT, ResNet and more, loadable in one line across TensorFlow, JAX and PyTorch backends.

Sarvam AI brings Indic-language text to speech, speech-to-text, translation and document parsing to Indian voice agents.

KcELECTRA-base-v2022: Korean ELECTRA pretrained on 162M Naver News comments for noisy Korean NLP

KcBERT: a free Korean BERT pretrained on 12.5GB of Naver news comments for messy, slang-heavy text.