The Known Good Updated 25 Jul 2026
gpt-4.5-preview vs OpenAI: GPT-5 Mini ·OpenAI vs OpenAI

gpt-4.5-preview vs OpenAI: GPT-5 Mini — head-to-head comparison

gpt-4.5-preview profile OpenAI: GPT-5 Mini profile All comparisons
gpt-4.5-preview — OpenAI
Quality
#43 of 61
47based on 4 of 6 evaluations
Known Good Index
Speed
unranked
Output tokens per second
Price
unranked
Blended per 1M tokens
Cache hit price
unranked
per 1M tokens
Context
unranked
tokens
OpenAI: GPT-5 Mini — OpenAI
Quality
#34 of 61
60based on 5 of 6 evaluations
Known Good Index
Speed
#47 of 316
99.9
Output tokens per second
Price
#162 of 340
$0.69
Blended per 1M tokens
In $0.25 · Out $2.00
Cache hit price
#39 of 206
$0.025
per 1M tokens (−90%)
Context
#104 of 345
400k
tokens (~600 pages)
Summary

On the Known Good Index, gpt-4.5-preview scores 47 and OpenAI: GPT-5 Mini scores 60 — a 13-point lead for OpenAI: GPT-5 Mini. They share 4 published evaluations, with gpt-4.5-preview ahead on 0 and OpenAI: GPT-5 Mini ahead on 4.

Published pricing is tracked for OpenAI: GPT-5 Mini only, at $0.69 per 1M tokens blended.

gpt-4.5-preview is proprietary; OpenAI: GPT-5 Mini is proprietary. 2 tracked hosts serve one or both.

Generated from tracked data on 25 Jul 2026 · how these comparisons are computed →

At a glance
Metricgpt-4.5-previewOpenAI: GPT-5 Mini
Known Good Index 47 based on 4 of 6 evaluations 60 based on 5 of 6 evaluations
Blended $/1M $0.69
Output tok/s 100
Context window 400k
Providers 0 2
Jump to section

Quality

Known Good Index and the capability indices, side by side

IndicesComponents
Index scores

Each index is min-max normalised across tracked models and scaled 0–100 from published evaluations. We do not run these evaluations.

Metricgpt-4.5-previewOpenAI: GPT-5 Mini
Known Good Index 47 60
Coding Index 50
Mathematics Index 59 93
Science Index 68 76
Reasoning Index 48 57
Agentic Index 50
Openness Index 5 5
The Known Good

Quality Breakdown

Every evaluation both models report a published score for

Evaluations both models report

Bar = gpt-4.5-preview · marker = OpenAI: GPT-5 Mini · scores ingested from Epoch AI (CC-BY 4.0)

GPQA Diamond69 / 75
MATH / AIME38 / 87
Humanity's Last Exam5 / 19
MATH Level 579 / 98

Right-hand figures read gpt-4.5-preview / OpenAI: GPT-5 Mini. gpt-4.5-preview leads on 0 of 4, OpenAI: GPT-5 Mini on 4.

The Known Good

Arena Elo

Human preference rating from blind pairwise votes

Arena ratings

Elo from blind pairwise votes, ingested from Arena. Higher is better.

Neither model currently appears in an arena we ingest.

The Known Good

Price & Cost

Published API pricing by token type

Per token typeCache discount
Published pricing

USD per 1M tokens. Blended is a 3:1 input:output mix. Live from OpenRouter, cross-checked against models.dev. Lower is better.

Metricgpt-4.5-previewOpenAI: GPT-5 Mini
Input $/1M $0.25
Output $/1M $2.00
Blended $/1M $0.69
Cache read $/1M $0.025
The Known Good
We publish blended price rather than cost-per-task. Cost-per-task requires token counts from an evaluation suite we do not run. Blended price is computed from published rates and is fully reproducible — see the pricing methodology.

Context Window

Maximum input accepted and maximum output emitted

Context and output limits

Provider-declared limits, from metadata via models.dev and OpenRouter. Higher is better.

Metricgpt-4.5-previewOpenAI: GPT-5 Mini
Context window 400k
Max output tokens 128k
The Known Good

Speed & Latency

Median throughput and time to first token across hosting providers

Output speedTime to first token
Throughput and latency

Medians across all tracked providers, derived from OpenRouter endpoint telemetry.

Metricgpt-4.5-previewOpenAI: GPT-5 Mini
Output tokens/s 100
Time to first token 3.3s
The Known Good

Providers

Which hosts serve each model, and at what price and speed

2 hosts
Full provider benchmarks →
Provider gpt-4.5-preview blendedgpt-4.5-preview tok/s OpenAI: GPT-5 Mini blendedOpenAI: GPT-5 Mini tok/s
OpenAI · first party $0.34 142
Azure · first party $0.69 78

Per-host pricing and telemetry from OpenRouter. A dash means that host does not serve that model.

The Known Good
Sources for this page
Pricing & providers — OpenRouter Metadata — models.dev Benchmarks — Epoch AI (CC-BY 4.0) Full attribution →