Comparisons of Medium Open Source AI Models (40B-150B)
Open source AI models with between 40 to 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.
gpt-oss-120B (high) and 
GLM-4.5-Air are the highest intelligence Medium open source models, defined as those with 40B-150B parameters, followed by 
Llama Nemotron Super 49B v1.5 (Reasoning) & GLM-4.5V (Reasoning).
Highlights
Intelligence
Artificial Analysis Intelligence Index; Higher is better
Total Parameters
Trainable parameters in billions
Further details
| Weights | Provider Benchmarks | |||||||
|---|---|---|---|---|---|---|---|---|
gpt-oss-120B OpenAI | 61 | 117B (5.1B active at inference time) | 131k | $0.3 | 216 | 🤗 |    +12 more | View |
GLM-4.5-Air Z AI | 53 | - | 128k | $0.4 | 150 | 🤗 |    +2 more | View |
Llama Nemotron Super 49B v1.5 (Reasoning) NVIDIA | 52 | 49B | 128k | - | - | 🤗 | ? | View |
GLM-4.5V (Reasoning) Z AI | 43 | 108B (12.0B active at inference time) | 64.0k | $0.9 | 95 | 🤗 |  | View |
Llama 3.3 Nemotron Super 49B v1 (Reasoning) NVIDIA | 39 | 49B | 128k | - | - | 🤗 | - | View |
 DeepSeek R1 Distill Llama 70B DeepSeek | 34 | 70B | 128k | $0.8 | 70 | 🤗 |   +5 more | View |
 Llama 4 Scout Meta | 33 | 109B (17B active at inference time) | 10.0M | $0.2 | 112 | 🤗 |  +11 more | View |
 Command A Cohere | 32 | 111B | 256k | $4.4 | 131 | 🤗 |  | View |
 Qwen2.5 Instruct 72B Alibaba | 31 | 72B | 131k | - | - | 🤗 |  +3 more | View |
Llama 3.3 Instruct 70B Meta | 31 | 70B | 128k | $0.6 | 118 | 🤗 |   +18 more | View |
Llama Nemotron Super 49B v1.5 NVIDIA | 30 | 49B | 128k | - | - | 🤗 | ? | View |
Llama 3.3 Nemotron Super 49B v1 NVIDIA | 29 | 49B | 128k | - | - | 🤗 | - | View |
 Mistral Large 2 (Nov '24) Mistral | 29 | 123B | 128k | $3.0 | 45 | 🤗 | | View |
Pixtral Large Mistral | 28 | 124B | 128k | $3.0 | 30 | 🤗 | | View |
Llama 3.1 Nemotron Instruct 70B NVIDIA | 26 | 70B | 128k | $0.2 | 27 | 🤗 | | View |
Llama 3.1 Instruct 70B Meta | 26 | 70B | 128k | $0.8 | 51 | 🤗 |   +10 more | View |
Mistral Large 2 (Jul '24) Mistral | 26 | 123B | 128k | $3.0 | 71 | 🤗 | | View |
Llama 3.2 Instruct 90B (Vision) Meta | 22 | 90B | 128k | $0.7 | 34 | 🤗 |  +1 more | View |
Qwen2 Instruct 72B Alibaba | 21 | 72B | 131k | - | - | 🤗 | | View |
 Hermes 3 - Llama-3.1 70B Nous Research | 17 | 70.6B | 128k | $0.1 | 23 | 🤗 | | View |
Llama 3 Instruct 70B Meta | 16 | 70B | 8.19k | $0.8 | 52 | 🤗 |  ? +7 more | View |
Mixtral 8x22B Instruct Mistral | 14 | 141B (39B active at inference time) | 65.4k | $3.0 | 53 | 🤗 | | View |
Qwen1.5 Chat 110B Alibaba | 13 | 110B | 32.0k | - | - | 🤗 | | View |
Qwen Chat 72B Alibaba | 10 | 72B | 33.8k | $1.0 | - | 🤗 | | View |
Command-R+ (Aug '24) Cohere | 9 | 104B | 128k | $4.4 | 38 | 🤗 | | View |
Llama 2 Chat 70B Meta | 8 | 70B | 4.10k | - | - | 🤗 | - | View |
Command-R+ (Apr '24) Cohere | 8 | 104B | 128k | $6.0 | 32 | 🤗 | | View |
 DBRX Instruct Databricks | 8 | 132B (36B active at inference time) | 32.8k | - | - | 🤗 | - | View |
 Jamba 1.5 Mini AI21 Labs | 6 | 52B (12B active at inference time) | 256k | $0.3 | - | 🤗 | | View |
Jamba 1.7 Mini AI21 Labs | 6 | 52B (12.0B active at inference time) | 258k | $0.3 | 137 | 🤗 | | View |
Jamba 1.6 Mini AI21 Labs | 5 | 52B (12B active at inference time) | 256k | $0.3 | 138 | 🤗 | | View |
Mixtral 8x7B Instruct Mistral | 5 | 46.7B (12.9B active at inference time) | 32.8k | $0.7 | 46 | 🤗 | ? +2 more | View |