ModelPriceWatch.com
Last scan 2026-08-28 Models tracked 237 Providers 31 Cheapest paid Granite 4.0 H Micro $0.017/Mtok in Every price links to its source

Solar Pro 4 vs GPT-5 mini

Side-by-side comparison of API pricing, specs, benchmarks, and capabilities

Solar Pro 4 is 92.4% cheaper on blended cost ($0.052 vs $0.688/Mtok)
Specification
Solar Pro 4Intro price
Overview
StatusCurrent mid tier Current mid tier
Released Aug 11, 2026 Aug 7, 2025
Pricing per million tokens
Input $0.030/Mtok $0.250/Mtok
Output $0.120/Mtok $2.00/Mtok
Blended avg $0.052/Mtok $0.688/Mtok
Cached input $0.006/Mtok $0.025/Mtok
Price basis Launch/intro pricing, per 1M tokens. Upstage's own rate card (upstage.ai/pricing/api) prints BOTH prices side by side for this SKU with the terms "90% off until 2026-09-10 (UTC). Standard pricing applies from 2026-09-11 (UTC)": input $0.30 -> $0.03, cached input $0.06 -> $0.006, output $1.20 -> $0.12. The tracked figures are the CURRENT EFFECTIVE discounted rates; on 2026-09-11 this record reverts to the $0.30/$0.06/$1.20 standard list price — a tracked price increase, watched by promo_until. Re-verified 2026-08-13: the rate card still shows the discount live and Upstage has NOT extended it. The page carries the discount as a dated phase table ("Solar Pro 4 — Promotional Pricing Phases", v1.2 of 2026-08-06) whose discount phase ends 2026-09-11T00:00:00Z and whose standard prices $0.30/$0.06/$1.20 are the ones baked into the card markup, so the vendor's own fallback when the promo lapses IS the standard rate. Upstage's launch post adds that the same discount runs on "Upstage Console and OpenRouter", which is why the gateway number also sits 10x under the standard list price and must not be read as one. Upstage lists all prices exclusive of 10% VAT.
Specifications
Context window 524K tokens 400K tokens
Parameters Proprietary Proprietary
Speed (TPS)
Modalities
Input
text
textimage
Benchmarks sources: Artificial Analysis, LMSYS Chatbot Arena (UC Berkeley), Upstage Solar Pro 4 launch blog / Artificial Analysis
Independent composite scores measured for both models — each on its own scale
AA Intelligence Index 42 25.8
AA Agentic Index 19.6
AA-Omniscience Index (−100–100) -17.3
LMSYS Chatbot Arena 1376 ELO
Per-benchmark results published for Solar Pro 4 only — no independent per-benchmark scores exist for GPT-5 mini yet
Avg benchmark score 57
Perf / dollar 1085.7
GPQA Diamond 89 vendor
SWE-Bench Verified 70.6 vendor
Terminal-Bench 2.1 57
LiveCodeBench 87.8 vendor
MCP Atlas 61.4 vendor
BrowseComp 49.2 vendor
Providers
Available from
Upstage — $0.030/$0.120/Mtok
OpenAI — $0.250/$2.00/Mtok

Cost at scale

1M tokens · 50/50 input/output
Projected cost of Solar Pro 4 vs GPT-5 mini at increasing token volumes
VolumeSolar Pro 4GPT-5 miniSavings
1M tokens $0.05 $0.69 $0.64 (93.1%)
10M tokens $0.53 $6.88 $6.35 (92.4%)
100M tokens $5.25 $68.75 $63.5 (92.4%)
1000M tokens $52.5 $687.5 $635 (92.4%)

When to pick which

distilled from the pricing and spec data above
Pick Solar Pro 4 if…
  • cost dominates: $0.052/Mtok blended vs $0.688 — 92.4% less on the same 50/50 token mix
  • you need the longer context: 524K tokens vs 400K (1.3×)
Pick GPT-5 mini if…
  • your workload re-reads context (agents, RAG, long chats): cached input costs $0.025/Mtok — 90% off its list input price

Summary

Solar Pro 4 by Upstage costs $0.030/Mtok input and $0.120/Mtok output, with a 524K-token context window. It supports text input.

GPT-5 mini by OpenAI costs $0.250/Mtok input and $2.00/Mtok output, with a 400K-token context window. It supports text, image input.

On a blended cost basis, Solar Pro 4 is 92.4% cheaper than GPT-5 mini.It also has a larger context window.

The two aren't directly comparable on average benchmark score: Solar Pro 4 has published per-benchmark results, while GPT-5 mini does not yet — it is measured today on independent composites (see the table above).

Note: Pricing is per million tokens. Actual costs vary with usage patterns, prompt caching, and batch discounts. Always verify against official provider pricing pages.

More comparisons

pairs sharing a model with this page