Skip to content

Open Source AI Models

Last updated: 43m ago

Ranked directory of 153 open source AI models with open weights you can self-host, fine-tune, and deploy without vendor lock-in. Scores computed from capabilities, pricing, context window, recency, and output capacity. Updated hourly.

Open Source Models

153

Average Score

49

Free to Use

23

Open Source Model Rankings

All 153 open source models ranked by composite score. Click any model for detailed benchmarks and analysis.

#ModelScore
1Gemma 4 26B A4B (free)Free692Gemma 4 31B (free)Free693Nemotron 3 Nano Omni (free)Free684MiMo-V2.5665Inkling666MiniMax M3657Qwen3.6 27B648Step 3.7 Flash639Qwen3.6 35B A3B6310Qwen3.5-35B-A3B6311Gemma 4 31B6312Qwen3.5-9B6313Nemotron 3 Super (free)Free6314Nemotron Nano 12B 2 VL (free)Free6315Qwen3.5-122B-A10B6316Qwen3.5 397B A17B6317Qwen3.5-27B6318Kimi K2.7 Code6219Kimi K2.66220Kimi K2.56221Gemma 4 26B A4B 6122Inkling Small6123Nemotron 3 Ultra (free)Free6024Kimi K35925GLM 4.6V5926North Mini Code (free)Free5927Nemotron 3.5 Content Safety (free)Free5828Laguna S 2.1 (free)Free5829Laguna XS 2.1 (free)Free5830DeepSeek V4 Pro5831DeepSeek V4 Flash 07315732DeepSeek V4 Flash 04235733MiniMax M3 (batch)5734MiMo-V2.5-Pro5735Qwen3 VL 30B A3B Thinking5736GLM 5.25737Qwen3 VL 8B Thinking5638Qwen3 VL 235B A22B Thinking5639Falcon-H1-Arabic 34B InstructFree5640Falcon-H1-Arabic 7B InstructFree5641Trinity Large Thinking5642gpt-oss-20b (free)Free5643GLM 5.15644GLM 55645MiniMax M2.55546Falcon-H1-Arabic 3B InstructFree5547Hy35548MiniMax M2.75549Nemotron 3 Super5550Reka Edge5551Kimi K2.7 Code (batch)5552Mistral Small 45453Nemotron 3 Nano 30B A3B5454GLM 4.75455LongCat 2.05456Inkling (batch)5457MiniMax M2.15458GLM 4.5V5359Laguna S 2.15360Qwen3 VL 8B Instruct5361Kimi K2 Thinking5362GLM 4.7 Flash5363Qwen3 VL 235B A22B Instruct5364Qwen3 VL 32B Instruct5365DeepSeek V3.25266Qwen3 VL 30B A3B Instruct5267Qwen3 Coder Next5268MiniMax M25269Qwen3 Next 80B A3B Thinking5270GLM 4.65271gpt-oss-safeguard-20b5172Nemotron Nano 9B V2 (free)Free5173Nemotron 3 Nano 30B A3B (free)Free5174Granite 4.1 8B5175Step 3.5 Flash5176Laguna XS 2.15177DeepSeek V3.2 Exp5178Ling-3.0-flash5179Nemotron 3 Ultra5080DeepSeek V3.1 Terminus5081GLM 5.2 (batch)5082Nemotron 3 Ultra (batch)5083gpt-oss-120b4984gpt-oss-20b4985GLM 4.54986Mistral Small 3.2 24B4987DeepSeek V3.14988Qwen3 30B A3B Thinking 25074989Llama 4 Scout4990Llama 4 Maverick4991Ministral 3 14B 25124992Ministral 3 8B 25124993Gemma 3 27B4894Ministral 3 3B 25124895Kimi K2 09054896Falcon Arabic 7B InstructFree4897Olmo 3 32B Think4898R1 05284799Qwen3 Next 80B A3B Instruct46100Jamba Large 1.746101Qwen3 Coder 480B A35B46102Gemma 3 12B46103Llama Guard 4 12B46104GLM 4.5 Air45105ERNIE 4.5 VL 424B A47B 45106Qwen3 Coder 30B A3B Instruct45107Qwen3 30B A3B Instruct 250745108Qwen3 30B A3B45109Hy3 preview45110Qwen3 32B45111Hunyuan A13B Instruct45112Qwen3 235B A22B45113Qwen3 14B45114Qwen3 235B A22B Instruct 250745115Qwen3 8B45116UI-TARS 7B 44117Granite 4.0 Micro44118Voxtral Small 24B 250744119Qwen3 235B A22B Thinking 250744120Falcon3 10B InstructFree43121Falcon3 7B InstructFree43122R143123MiniMax-0143124Gemma 3 4B42125DeepSeek V3 032442126Kimi K2 071142127Reka Flash 341128Mistral Small 3.1 24B40129Command A39130DeepSeek V339131Stable Diffusion 3.539132Llama 3.3 70B Instruct38133Qwen2.5 VL 72B Instruct36134Llama 3.1 8B Instruct36135Qwen2.5 7B Instruct35136ALLaM 7B Instruct (preview)Free35137Llama 3.1 70B Instruct34138Mistral Nemo34139Mistral Small 334140Qwen2.5 72B Instruct34141Llama 3.2 3B Instruct34142Phi 433143Falcon Mamba 7B InstructFree33144Gemma 3n 4B32145R1 Distill Llama 70B32146Mixtral 8x22B Instruct32147WizardLM-2 8x22B30148Qwen2.5 Coder 32B Instruct29149Llama 3.2 1B Instruct29150Gemma 2 27B26151LTX-Video 2Free20152Wan 2.1 T2VFree18153Stable Video Diffusion6

Top Open Source Providers

Which companies and organizations contribute the most open source AI models.

Alibaba

32 open source models

DeepSeek

12 open source models

NVIDIA

11 open source models

Zhipu AI

11 open source models

Mistral AI

10 open source models

Google

9 open source models

Moonshot AI

8 open source models

Meta

8 open source models

Why Open Source AI Models?

Open source AI models give you full control over your AI stack. Here is why teams choose open weights over proprietary APIs.

Data Privacy & Control

Self-host models on your own infrastructure. Your data never leaves your network, meeting strict compliance and privacy requirements.

No Vendor Lock-in

Open weights mean you can deploy anywhere: on-premises, any cloud provider, or at the edge. Switch infrastructure without changing your model.

Customization & Fine-tuning

Fine-tune open source models on your proprietary data. Adapt architecture, optimize inference, and build domain-specific solutions.

Community-Driven Innovation

Benefit from global research communities. Open models receive rapid improvements, safety patches, and community-built tooling.

Open Source vs Proprietary AI Models

AspectOpen SourceProprietary
Data PrivacyFull control, self-hostedData sent to provider
CustomizationFine-tune, modify weightsPrompt engineering only
Cost at ScaleFixed infrastructure costPer-token pricing
Setup EffortRequires infrastructureAPI key and go
Latest CapabilitiesCommunity-driven releasesCutting-edge, provider-managed
SupportCommunity forums, docsEnterprise SLAs available
Frequently Asked Questions

The top open source AI models in 2026 include Meta's Llama 4 family, DeepSeek R1 and V3, Mistral's Large and Medium models, Alibaba's Qwen 2.5 series, and Google's Gemma. These models offer competitive performance with full access to model weights, enabling self-hosting and fine-tuning. Rankings are updated hourly based on benchmarks, capabilities, and community adoption.

Open source AI models are free to download and self-host - you only pay for your own compute infrastructure (GPU servers). Many providers also offer free API access to popular open source models through various API providers. Self-hosting eliminates per-token API costs entirely, which can save significant money at scale.

Yes, the performance gap has narrowed significantly. Models like DeepSeek R1, Llama 4 Maverick, and Qwen 2.5 now match or exceed GPT-4 on many benchmarks including coding, math, and reasoning tasks. While frontier proprietary models still lead on the most complex tasks, open source alternatives are viable for the vast majority of production use cases.

The most popular tools for running open source models locally are Ollama (easiest setup, one-line install), llama.cpp (optimized C++ inference for consumer hardware), and vLLM (high-throughput production serving). For consumer GPUs, quantized versions (GGUF format) let you run 7B-70B parameter models on hardware with 8-48GB VRAM. Cloud GPU providers like RunPod and Lambda also offer on-demand hosting.

Open weight models (like Llama and Gemma) release the trained model weights so you can run and fine-tune them, but may not release the full training code, datasets, or training infrastructure. Truly open source models release everything - weights, training code, data pipelines, and evaluation suites - under permissive licenses. Most models marketed as "open source" are technically open weight, which still provides the key benefits of self-hosting and customization.

Explore More AI Model Data

Dive deeper into AI model rankings, pricing breakdowns, and head-to-head comparisons across all providers.

Related Guides

Best Open Source AI Models & LLM Leaderboard (2026) | LM Market Cap