Sulat.com
AI models
$10 off the fastest DeepSeek V4.1 Flash, Kimi K3 and GLM 5.3 from Synthetic
Upstage logo

Model details

solar-pro3

Solar Pro 3 is a flagship language model built on a Mixture-of-Experts architecture, featuring 102 billion parameters. Designed to serve as a versatile tool for complex language tasks, it excels at generating, summarizing, and understanding text. Its design intent focuses on providing high-level reasoning capabilities, making it a strong choice for users who need to navigate intricate logic and chain-of-thought processes in their workflows.

This model represents a significant evolution in the Solar Pro family, building upon previous generations to offer enhanced performance and multilingual support. By leveraging its large-scale expert-based design, it is optimized to handle demanding analytical requirements while maintaining efficiency. Its practical strengths lie in its ability to act as a specialized agent for reasoning-heavy applications, positioning it as a forward-looking solution for developers seeking to integrate sophisticated logic into their text-based systems.

Upstagesolar-pro3solar-pro

Quick Info

Powered by
Provider
Upstage
Model key
solar-pro3
Release date
Jan 1, 2026
Last updated
Jan 1, 2026
Knowledge cutoff
2025-03
Input modalities
Output modalities
Capabilities

Cost

Input token cost
$0.25
Output token cost
$0.25

Limits

Output tokens
8,192 tokens
Context window
131,072 tokens

Transparent token rates

Compare solar-pro3 pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about solar-pro3

Upstage

Official sourceBenchmark

Upstage has published a first-party blog post announcing Solar Pro 3 as the March 2026 update to its Solar Pro large language model line. The post states that Solar Pro 3 delivers roughly 2× gains over Solar Pro 2 across every evaluation category, with the headline improvement being agentic performance — Tau2-all rises According to the same announcement, the agentic improvements are driven by SnapPO, Upstage's proprietary in-house reinforcement learning framework developed through the Solar Open model research program, which is designed to compose independent training stages so that reasoning can be improved simultaneously across mat

Videos about solar-pro3

More models around solar-pro3