ModelBeat
Models/Solar Pro 4
All models
U
upstage/solar-pro4

Solar Pro 4 is Upstage’s AI model, released Aug 10, 2026. It costs $0.30/M input and $1.20/M output tokens, and handles a 524K-token context window.

Solar Pro 4 is Upstage's cost-efficient large language model, featuring a 524K context window.

Context
524K
Input
$0.30 / 1M
Output
$1.20 / 1M
Cheapest via KT
$0.30 / $1.20 / 1M
Released
Aug 10, 2026
Modalities
text→text

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Solar Pro 4 lands among all models tracked on Model Beat.

44.0
Intelligence Index
Artificial Analysis
44th percentile of tracked models
41.0
Coding Index
Artificial Analysis
41st percentile of tracked models

Reasoning

2 evals
GPQA Diamond89.1%

Graduate-level scientific reasoning

Humanity's Last Exam29.2%

Frontier of human expert knowledge

Coding

1 evals
SciCode44.6%

Scientific research coding

Changelog

3
  • Solar Pro 4: cheapest credible provider via KT now $0.3 per 1M input tokens (was $0.03, +900%)cheapest provider · Nov 12, 2026
  • Solar Pro 4: cheapest credible provider via KT now $1.2 per 1M output tokens (was $0.12, +900%)cheapest provider · Nov 12, 2026
  • Solar Pro 4 (Upstage) added to the tracker: $0.03/$0.12 per 1M, 524K context.tracked · Aug 11, 2026

Compare Solar Pro 4 with

Frequently asked questions

What is Solar Pro 4?

Solar Pro 4 is Upstage's cost-efficient large language model, featuring a 524K context window.

Who created Solar Pro 4?

Solar Pro 4 is an AI model developed by Upstage, released Aug 10, 2026.

How much does Solar Pro 4 cost?

Solar Pro 4's list price from Upstage is $0.30 per million input tokens and $1.20 per million output tokens. The cheapest credible third-party provider on OpenRouter (KT) serves it at $0.30/$1.20 per 1M.

What is Solar Pro 4's context window?

Solar Pro 4 supports a context window of up to 524K tokens.

How does Solar Pro 4 perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Solar Pro 4 scores 89.1% on GPQA Diamond, 29.2% on Humanity's Last Exam, 44.6% on SciCode.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Nov 19, 2026.