Comparisons of Large Open Source AI Models (>150B)
Open source AI models with over 150 billion parameters. Models are considered Open Source (also commonly referred to as open weights) where their weights are accessible to download. This allows self-hosting on your own infrastructure and enables customizing the model such as through fine-tuning. Click on any model to see detailed metrics. For more details including relating to our methodology, see our FAQs.
Qwen3 235B 2507 (Reasoning) and 
DeepSeek R1 0528 are the highest intelligence Large open source models, defined as those with >150B parameters, followed by 
GLM-4.5 & 
MiniMax M1 80k.
Highlights
Intelligence
Artificial Analysis Intelligence Index; Higher is better
Loading chart...
Total Parameters
Trainable parameters in billions
Loading chart...
Further details
| Weights | Provider Benchmarks | |||||||
|---|---|---|---|---|---|---|---|---|
Qwen3 235B A22B 2507 (Reasoning) Alibaba | 64 | 235B (22B active at inference time) | 256k | $2.6 | 65 | Not available |    +3 more | View |
DeepSeek R1 0528 (May '25) DeepSeek | 59 | 685B (37B active at inference time) | 128k | $1.0 | 24 | 🤗 |    +12 more | View |
GLM-4.5 Z AI | 56 | 355B (32B active at inference time) | 128k | $1.0 | 62 | 🤗 |   +3 more | View |
MiniMax M1 80k MiniMax | 53 | 456B (45.9B active at inference time) | 1.00M | $0.8 | - | 🤗 | | View |
MiniMax M1 40k MiniMax | 51 | 456B (45.9B active at inference time) | 1.00M | $0.8 | - | 🤗 | | View |
Qwen3 235B A22B 2507 (Non-reasoning) Alibaba | 51 | 235B (22.0B active at inference time) | 256k | $1.2 | 41 | 🤗 | +5 more | View |
DeepSeek R1 (Jan '25) DeepSeek | 50 | 685B (37B active at inference time) | 128k | $2.0 | - | 🤗 |  +10 more | View |
 Kimi K2 Moonshot AI | 49 | 1.0KB (32B active at inference time) | 128k | $1.1 | 45 | 🤗 |  +6 more | View |
Qwen3 235B A22B (Reasoning) Alibaba | 48 | 235B (22B active at inference time) | 32.8k | $2.6 | 35 | 🤗 |  +5 more | View |
 Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) NVIDIA | 46 | 253B | 128k | $0.9 | 43 | 🤗 | | View |
Qwen3 Coder 480B A35B Alibaba | 45 | 480B (35.0B active at inference time) | 262k | $3.0 | 49 | 🤗 | +7 more | View |
DeepSeek V3 0324 (Mar '25) DeepSeek | 44 | 671B (37B active at inference time) | 128k | $0.5 | 22 | 🤗 |  +10 more | View |
 Llama 4 Maverick Meta | 42 | 402B (17B active at inference time) | 1.00M | $0.4 | 166 | 🤗 |  +11 more | View |
DeepSeek V3 (Dec '24) DeepSeek | 35 | 671B (37B active at inference time) | 128k | $0.5 | - | 🤗 | +6 more | View |
Qwen3 235B A22B Alibaba | 33 | 235B (22B active at inference time) | 32.8k | $1.2 | 36 | 🤗 | +4 more | View |
Llama 3.1 Instruct 405B Meta | 29 | 405B | 128k | $3.3 | 35 | 🤗 |  +11 more | View |
MiniMax-Text-01 MiniMax | 29 | 456B (45.9B active at inference time) | 4.00M | $0.4 | 32 | 🤗 | | View |
 Llama 3.1 Tulu3 405B Allen Institute for AI | 29 | 405B | 128k | - | - | 🤗 | - | View |
DeepSeek-V2.5 (Dec '24) DeepSeek | 24 | 236B (21B active at inference time) | 128k | $0.2 | - | 🤗 | | View |
DeepSeek-V2.5 DeepSeek | 23 | 236B (21B active at inference time) | 128k | $0.2 | - | 🤗 | | View |
 R1 1776 Perplexity | 22 | 671B (37B active at inference time) | 128k | $3.5 | - | 🤗 | | View |
 Grok-1 xAI | 21 | 314B (78B active at inference time) | 8.19k | - | - | 🤗 | - | View |
 Jamba 1.7 Large AI21 Labs | 18 | 398B (94.0B active at inference time) | 256k | $3.5 | 50 | 🤗 | | View |
Jamba 1.5 Large AI21 Labs | 18 | 398B (94B active at inference time) | 256k | $3.5 | - | 🤗 | | View |
DeepSeek-Coder-V2 DeepSeek | 17 | 236B (21B active at inference time) | 128k | $0.2 | - | 🤗 | | View |
Jamba 1.6 Large AI21 Labs | 17 | 398B (94B active at inference time) | 256k | $3.5 | 49 | 🤗 | | View |
DeepSeek-V2-Chat DeepSeek | 11 | 236B (21B active at inference time) | 128k | $0.2 | - | 🤗 | | View |
 Arctic Instruct Snowflake | 10 | 480B (17B active at inference time) | 4.00k | - | - | 🤗 | - | View |