VerdictPal
Join

Models / OpenAI

GPT-5 (medium)

GPT-5 · Released 2025-08-07

AA Intelligence Index 33.7. Listed on Artificial Analysis with API pricing $1.25/$10 per 1M tokens. Benchmark rows sync from the AA snapshot — see the ledger on this page.

Frontier model tracked on Artificial Analysis. Use the benchmark ledger below for sourced scores; we do not infer numbers beyond published rows.

#99 of 184 on AA Index · current snapshot
23 AA Index · 12 benchmark rows · 1 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite65
ContextUnlisted
Input / 1M$1.3
Output / 1M$10
Knowledge cutoffunpublished
Statussolid

Family profile

Best published score in each covered benchmark family.

76/100 avg
Knowledge
87
Reasoning
84
Math
99
Coding
70
Agentic
38
Long context
76

6 tested benchmark families

Benchmark placements

Where GPT-5 (medium) places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
MATHMath#3/ 2599Artificial Analysis2026-09-10
MMLU-ProKnowledge#14/ 3987Artificial Analysis2026-09-10
AIMEMath#16/ 3792Artificial Analysis2026-09-10
LiveCodeBenchCoding#23/ 3170Artificial Analysis2026-09-10
IFBenchReasoning#51/ 12771Artificial Analysis2026-09-10
AA-LCRLong context#76/ 18276Artificial Analysis2026-09-10
Humanity's Last ExamReasoning#104/ 18525Artificial Analysis2026-09-10
Artificial Analysis Intelligence IndexReasoning#108/ 18823Artificial Analysis2026-09-10
GPQA DiamondReasoning#110/ 19084Artificial Analysis2026-09-10
Terminal-BenchAgentic#125/ 18938Artificial Analysis2026-09-10
AA output speedPerformance#127/ 190Artificial Analysis2026-09-10
AA time to first tokenPerformance#127/ 190Artificial Analysis2026-09-10
Family context

The three highest-scoring models with pages in each capability family. Where GPT-5 (medium) shows up, it's highlighted.

KnowledgeGPT-5 (medium) · 87
  1. 1
    GPT 5.5
    91
  2. 2
    Claude Opus 4.8
    91
  3. 3
    Claude Sonnet 4.6
    91
ReasoningGPT-5 (medium) · 84
  1. 1
    GPT-6 Astra
    96
  2. 2
    Claude Mythos Preview
    95
  3. 3
    Gemini 3.7 Flash
    95
MathGPT-5 (medium) · 99
  1. 1
    OpenAI o3
    99
  2. 2
    GPT-5 (medium)
    99
  3. 3
    Grok 4
    99
CodingGPT-5 (medium) · 70
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticGPT-5 (medium) · 38
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextGPT-5 (medium) · 76
  1. 1
    Kimi K3
    89
  2. 2
    Claude Fable 5.1
    85
  3. 3
    GPT-5.5 (high)
    84
Newest receipts

Benchmark rows added to the public ledger for GPT-5 (medium) in the last 120 days. Older rows live in the full table below.

12 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context76%Artificial Analysis6
AA output speedPerformance0 t/sArtificial Analysis6
AA time to first tokenPerformance0.00sArtificial Analysis6
AIMEMath91.67%Artificial Analysis6
Artificial Analysis Intelligence IndexReasoning22.9%Artificial Analysis6
GPQA DiamondReasoning84.2%Artificial Analysis6
Humanity's Last ExamReasoning25.4%Artificial Analysis6
IFBenchReasoning70.61%Artificial Analysis6
LiveCodeBenchCoding70.3%Artificial Analysis6
MATHMath99.13%Artificial Analysis6
MMLU-ProKnowledge86.7%Artificial Analysis6
Terminal-BenchAgentic37.88%Artificial Analysis6
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Knowledge87
Reasoning84
Math99
Coding70
Agentic38
Long context76
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMetaxAIMiniMaxZhipuXiaomiOtherMeituanCohereNVIDIA

Full benchmark ledger

Every catalog benchmark for GPT-5 (medium). Scores link to the original source; gaps mean no public row exists yet.

12 sourced rows

12 of 37 catalog benchmarks have a sourced row for GPT-5 (medium).

Knowledge

Reasoning

Math

Coding

Agentic

Long context

Performance

Sources

The Pack · Editorial newsletter

New tools in your inbox. Free.

One short email when a tool ships or changes status. No tracking, no third-party analytics. Unsubscribe in one click.