Which open models can you run locally?
154 current open-weight models from 44 labs. Set your memory to see what fits.
Language models
All 47ModelLabParamsYour deviceScore
Kimi K3Moonshot AI2.78T1475 GB ◇92.0% GPQAMiniMax M3MiniMax427B797 GB ◇90.5% GPQADeepSeek V4 ProDeepSeek1.6T807 GB ◇90.1% GPQAHy4 previewTencent780B1455 GB ◇89.7% GPQADeepSeek V4.1 FlashDeepSeek763.2B476 GB ◇88.8% GPQAModels with a GPQA Diamond score come first, highest score first. The rest are newest first.
Coding
All 9ModelLabParamsYour deviceScore
colgrep agent 2BLightOn2.52B2.3 GB ◇Mellum2.1 ThinkingJetBrains12.1B7.5 GB ◇AesCodeMicrosoft8.77B18.0 GB ◇Qwen 27B BlendJetBrains27.8B14.3 GB ◇Nex N2.5Nex-AGI35.1B66.6 GB ◇Newest first.
Agents
All 5ModelLabParamsYour deviceScore
colgrep agent 2BLightOn2.52B2.3 GB ◇Nex N2.5Nex-AGI35.1B66.6 GB ◇UI-MateTencent9.41B19.1 GB ◇Fara 1.5Microsoft4.54B10.0 GB ◇Qwen AgentWorldAlibaba Qwen34.7B65.7 GB ◇Newest first.
Speech to text
All 10ModelLabParamsYour deviceScore
Voxtral Realtime ArabicMistral AI4.43B—VibeVoice ASR StreamingMicrosoft2.81B—Granite Speech 5.0 TurboCTCIBM473M—VibeVoice ASR BitNetMicrosoft323M—MOSS Transcribe DiarizeOpenMOSS909M—Newest first.
Text to speech
All 9ModelLabParamsYour deviceScore
AuKTencent——IndexTTS 2.5Bilibili IndexTeam——Higgs TTS 3Boson AI4.65B—Chatterbox FlashResemble AI——MOSS TTS v1.5OpenMOSS8.49B—Newest first.
Audio understanding and voice chat
All 8ModelLabParamsYour deviceScore
Nemotron 3 DiarizationNVIDIA99M—GLM-4 Voice of ReasonKyutai9.54B—NemotronLabs VoiceChatPreviewNVIDIA11.1B—Nemotron Labs AudexPreviewNVIDIA——Moshika RL SeamlessKyutai7.69B—Newest first.
Image generation
All 8ModelLabParamsYour deviceScore
Qwen Image 2.1 TurboAlibaba Qwen7.12B—PixelUMMNVIDIA——Qwen Image 2.1Alibaba Qwen7.12B—Ming Image 0.1Ant Group inclusionAI6.15B—LLaDA ImageAnt Group inclusionAI6.54B—Newest first.
Video and 3D
All 11ModelLabParamsYour deviceScore
MiniMax H3MiniMax33.1B—LTX 2.5Lightricks——Cosmos H DreamsNVIDIA——Wan 2.2 Animate 2Alibaba Wan——Wan DancerAlibaba Wan——Newest first.
Music and sound effects
All 5ModelLabParamsYour deviceScore
MiniMax Music3MiniMax2.43B—Magenta Realtime 2Google——MOSS SoundEffect v2OpenMOSS1.42B—Stable Audio 3Stability AI2.31B—ACE Step 1.5 XLACE-Step4.17B—Newest first.
Translation
All 5ModelLabParamsYour deviceScore
Index EchoBilibili IndexTeam—1.5 GB ◇Index HomuraBilibili IndexTeam2.27B1.8 GB ◇Index TranslateBilibili IndexTeam2.27B1.8 GB ◇North Small TranslatePreviewCohere218.3B407 GB ◇Hy MT2Tencent2.04B1.5 GB ◇Newest first.
OCR and document parsing
All 10ModelLabParamsYour deviceScore
Youtu Parsing OmniTencent5.33B—LightOnOCR 3LightOn1.01B—WeVisDocTencent2.44B—Jina OCR v1Jina AI3.37B—ArmorOCRAnt Group inclusionAI8.77B—Newest first.
Embedding and reranking
All 8ModelLabParamsYour deviceScore
EmbeddingGemma 2Google744M—EVIETencent4.54B—BitNet EmbeddingMicrosoft——Jina Reranker 3.5Jina AI597M—Nemotron 3 EmbedNVIDIA1.14B—Newest first.
Safety filters
All 4ModelLabParamsYour deviceScore
Shieldstral 1.0Mistral AI3.85B—SingGuardAnt Group inclusionAI1.11B—Privacy FilterOpenAI1.4B—Granite Guardian 4.1IBM8.38B—Newest first.
Specialized models
All 17ModelLabParamsYour deviceScore
LIDstral ArabicMistral AI——AREX 2BAAI27.4B—FLUX 3 ActionPreviewBlack Forest Labs——EXAONE ForecastLG AI Research——Recon2ReasonBAAI4.44B—Newest first.
Min memory is the smallest listed option. ● stated by the publisher · ◇ estimate from file size, an 8K-token context and 0.5 GB for the runtime. — means no figure is available.