Open-Weight Models
Open-weight models from Meta, Mistral, and DeepSeek for self-hosting and fine-tuning.
Open-weight models from Meta, Mistral, and DeepSeek for self-hosting and fine-tuning.
Models
14
in category
Providers
3
represented
Avg. API Cost
$2.02
combined in+out / 1M
Top Model
DeepSeek V4
Price
Open-Weight Models Rankings
Sorted by price
| # | Model | Provider | Context | Input / 1M |
|---|---|---|---|---|
| 1 | Ministral 8B | Mistral | 128K | $0.10 |
| 2 | Mistral Small | Mistral | 32K | $0.10 |
| 3 | DeepSeek Coder V2 | DeepSeek | 128K | $0.14 |
| 4 | Llama 3.3 70B | Meta | 128K | $0.23 |
| 5 | Llama 4 Scout | Meta | 10M | $0.15 |
| 6 | Llama 4 Maverick | Meta | 1M | $0.20 |
| 7 | Llama 3.1 405B | Meta | 128K | $0.35 |
| 8 | Codestral | Mistral | 256K | $0.30 |
| 9 | DeepSeek V3 0324 | DeepSeek | 128K | $0.27 |
| 10 | DeepSeek V3 | DeepSeek | 64K | $0.27 |
Related Models
DeepSeek V4
Next-gen DeepSeek general model emphasizing coding and cost efficiency.
Why: DeepSeek · Price leader
DeepSeek V3 0324
Updated V3 with improved instruction following and 128K context.
Why: DeepSeek · Price leader
Llama 4 Maverick
Open multimodal Llama with 1M context window.
Why: Meta · Price leader
Llama 4 Scout
Llama 4 with industry-leading 10M context for documents.
Why: Meta · Price leader
Mistral Small
Compact Mistral for cost-sensitive applications.
Why: Mistral · Price leader
DeepSeek R1
Open reasoning model with o1-class performance at fraction of cost.
Why: DeepSeek · Price leader
Codestral
Code-specialized Mistral model for IDE and agent use.
Why: Mistral · Price leader
DeepSeek V3
Highly efficient MoE model rivaling frontier closed models.
Why: DeepSeek · Price leader
Related Providers
Relevant Comparisons
DeepSeek V4 vs DeepSeek V3 0324
Why: Top models in this category
DeepSeek V4 vs Llama 4 Maverick
Why: Top models in this category
DeepSeek V4 vs Llama 4 Scout
Why: Top models in this category
DeepSeek V3 0324 vs Llama 4 Maverick
Why: Top models in this category
DeepSeek V3 0324 vs Llama 4 Scout
Why: Top models in this category
Llama 4 Maverick vs Llama 4 Scout
Why: Top models in this category
Calculators
Timeline Events
Frequently Asked Questions
- What is the open-weight models?
- Based on SGO's dataset, Llama 3.3 70B by Meta ranks highly for price with a price score of 84. Rankings are computed from structured benchmark and pricing data.
- How are open-weight models ranked?
- Rankings are derived from the centralized SGO model dataset — benchmark scores, API pricing, context windows, and capability flags. No recommendations are hardcoded.
- How many models match this category?
- 14 models in the SGO catalog currently match this category.
Explore More
14 models · Data derived from the SGO intelligence catalog