Solar Pro 4, the Upstage model in numbers
Solar Pro 4 in numbers: price per token, context window, independent benchmark scores and comparable models, based on public data checked on October 22, 2026.
Data as of October 22, 2026
Key figures
| Developer | Upstage |
|---|---|
| Released | August 6, 2026 |
| Knowledge cutoff | February 2026 |
| Context window | 524,288 tokens (≈ 786 pages) |
| Maximum output | 131,072 tokens |
| Input price | $0.30 per million tokens (developer's price) |
| Output price | $1.2 per million tokens |
| Open weights | No |
| Reasoning mode | Yes |
| Tool calling | Yes |
| Structured output (JSON) | Yes |
| Accepted inputs | text |
| Providers offering it | 6 |
What you need to know
Solar Pro 4 is a language model from Upstage, released on August 6, 2026. That makes it 2 months old as of October 22, 2026. It is a closed model: you access it through an API, from Upstage or one of the 6 providers listed on models.dev. It has a reasoning mode: it can "think" before answering, which helps with multi-step problems but uses more tokens, and therefore more money.
We file it under budget models (24 models in this category on the site).
Pricing and budget
Solar Pro 4 costs $0.30 per million input tokens and $1.2 per million output tokens (developer's price, checked on October 22, 2026).
To put it in perspective: reading 20,000 tokens (about thirty pages) and writing 2,000 costs roughly $0.008; a thousand of those tasks, $8.4.
That is in line with budget models, whose median output price is $1.13.
Tokens read back from cache cost $0.06 per million, which helps an agent that keeps resending the same context.
To estimate your own budget, use the agent cost calculator.
What do independent benchmarks say?
As of October 22, 2026, Epoch AI publishes no score for Solar Pro 4 on our basket's benchmarks. To place it, rely on its specifications and price, then test it on your own cases.
Context and capabilities
Its context window reaches 524,288 tokens, roughly 786 pages of text at 500 words per page: enough for a long report or several contracts at once. A single answer can run to 131,072 tokens.
It accepts as input: text. Its knowledge stops at February 2026: beyond that, it needs web search or your documents.
It can call tools (functions, APIs) and return structured JSON output, which makes it a possible foundation for an AI agent.
When to choose it
Given its category, Solar Pro 4 is mainly aimed at high volumes where every cent counts: sorting, information extraction, summaries, simple assistants.
Before adopting it, test it on your own examples: benchmarks measure standard tasks, not your job.
What could replace it?
In the same category, the models closest to Solar Pro 4 are:
- GPT-5.6 Luna (OpenAI, index 79.6)
- MAI-Code-1.1-Flash (Microsoft (Phi, MAI), $1.2 per million output tokens)
- GPT-5.4 nano (OpenAI, index 71.7)
- Qwen3.6 Flash (Alibaba Cloud (Qwen), index 64.5)
Similar models
- GPT-5.6 Luna · OpenAI · index 79.6
- MAI-Code-1.1-Flash · Microsoft (Phi, MAI)
- GPT-5.4 nano · OpenAI · index 71.7
- Qwen3.6 Flash · Alibaba Cloud (Qwen) · index 64.5
Other Upstage models
- Solar Mini 4 · September 22, 2026
Frequently asked questions
How much does Solar Pro 4 cost?
As of October 22, 2026, Solar Pro 4 costs $0.30 per million input tokens and $1.2 per million output tokens (developer's price). A task that reads 20,000 tokens and writes 2,000 costs about $0.008.
What is the context window of Solar Pro 4?
524,288 tokens, roughly 786 pages of text. An answer can reach 131,072 tokens.
Is Solar Pro 4 open source?
No. It is a closed Upstage model, available only through an API.
When was Solar Pro 4 released?
Solar Pro 4 was released on August 6, 2026, according to models.dev.
Can you build an agent with Solar Pro 4?
Yes: it can call tools and return structured JSON, the two basic building blocks of an agent.
Data sources
- Models.dev — Model data: models.dev (open-source project)
- OpenRouter — Market prices cross-checked with OpenRouter
Data as of October 22, 2026
Report an error
Does a figure, a date or a fact look wrong? Tell us: every report is checked.