BenchLeader

GPT-5

Reasoning effort
Quality#50 of 339
59.0Index
Top quarter
+8.8 on the median
Price#197 of 255
$3.44per 1M tokens
Dearest quarter
4.0× the median
Speed#120 of 218
74tokens/s
Below the median
1.0× the median
Value#195 of 255
17.2points per $
Bottom quarter
0.3× the median

Better value: GLM 5.3 Flash scores 57.8 for $0.238/M, 14× cheaper.

15 cheaper ranked models score higher: GPT-6.1 Terra (64.1, $2.00), Gemini 4 Hydrogen (63.4, $3.00), Gemini 3.5 Flash (62.5, $3.38) and 12 more.

As of 15 Nov 2026, GPT-5 by OpenAI scores 59.0 on the BenchLeader Index (608 ranked configurations), with its best results in long context (70) and weakest in reasoning (48). API pricing: $1.25 per 1M input tokens and $10.00 per 1M output tokens ($3.44 blended at 3:1), list price on 15 Nov 2026. It streams about 74 tokens/s (Artificial Analysis). Released 7 Aug 2025.

Key facts (as of 15 Nov 2026)

  • GPT-5 API pricing: $1.25 per 1M input tokens and $10.00 per 1M output tokens ($3.44 blended at 3:1) (list price, 15 Nov 2026).
  • GPT-5 streams about 74 tokens/s (Artificial Analysis, 15 Nov 2026).
  • GPT-5 was released on 7 Aug 2025 (OpenAI).

In detail. GPT-5 is an OpenAI proprietary reasoning model, released 7 Aug 2025. Its best configuration (medium reasoning effort) ranks #113 of 608 on the BenchLeader Index at 59.0 ±5.4, in the upper half. The ± is the point: 258 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in long context (70) and lowest in reasoning (48). At $3.44 per million tokens blended it is pricier than most ranked models. Output speed of 74 tokens per second puts it slower than most, with a first answer in 25.3 s. It has been measured at 6 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 11 Nov 2026.

Blended price
$3.44/M
$1.25 in · $10.00 out
via Artificial Analysis
Output speed
74 tok/s
measured by Artificial Analysis
First answer
25 s
first token 4.60 s
Context
400k
via models.dev
Full answer
66 s
last measured 13 Nov 2026
Cost per task
–
Tokens per task
–
Released
7 Aug 2025
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index59
  2. Reasoning48
  3. Coding57
  4. Agents & tools58
  5. Maths65
  6. Instruction following69
  7. Multimodal57
  8. Long context70

Embed this rank

More badge styles
[![GPT-5 on BenchLeader: #51 overall](https://www.benchleader.com/badge/gpt-5.svg?style=compact)](https://www.benchleader.com/models/gpt-5)

Paste it at the top of README.md, beside your other badges. The badge redraws with every data update and links back to this page.

Get an email when OpenAI ships a new model

And once when it gets ranked, plus the weekly review on Mondays.

Choose providers and emails

Versions

OpenAI has shipped 13 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.

ModelReleasedIndexRank
GPT-5.69 Jul 2026––
GPT-5.5xhigh23 Apr 202664.1#31
GPT-5.4xhigh5 Mar 202663.1#42
GPT-5.2xhigh11 Dec 202559.9#95
GPT-5.1high12 Nov 202557.5#151
GPT-5mediumthis page7 Aug 202559.0#113
GPT-4.114 Apr 202548.7#377
GPT-4.527 Feb 202550.5#335
GPT 4 012525 Jan 2024––
GPT-4 11066 Nov 2023––
GPT 4 061313 Jun 202339.5#541
GPT-414 Mar 202342.3#496
GPT-3.5–––

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.

EffortIndexRankSpeedFirst answerChat reply costCategories
minimal45.2#45865 tok/s2.76 s$0.0035Agents & tools 50 · Instruction following 48 · Maths 38 · Multimodal 44 · Reasoning 38
low53.5#25877 tok/s9.39 s$0.0035Agents & tools 57 · Coding 70 · Instruction following 66 · Maths 45 · Multimodal 56 · Reasoning 36
mediumbest59.0#11374 tok/s25 s$0.0035Agents & tools 58 · Coding 57 · Instruction following 69 · Long context 70 · Maths 65 · Multimodal 57 · Reasoning 48
high58.2#13480 tok/s55 s$0.0035Agents & tools 54 · Coding 60 · Human preference 62 · Instruction following 72 · Knowledge 58 · Long context 64 · Maths 57 · Multimodal 55 · Reasoning 49
thinking––69 tok/s4.60 s$0.0035Instruction following 59
not stated56.6#17469 tok/s4.60 s$0.0035Agents & tools 49 · Coding 49 · Human preference 60 · Instruction following 57 · Knowledge 62 · Long context 57 · Maths 66 · Multimodal 59 · Reasoning 59

How its index has moved

8 Sep 2026 to 15 Nov 2026
465667
  • medium
  • high
  • not stated
  • low

The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Metadata

Benchmarkminimallowmediumhighnot statedSourceTrend
AA Openness Indexnot in index5.6#3075.6#3075.6#3075.6#30711.1#292Artificial Analysis

Reasoning

Benchmarkminimallowmediumhighnot statedSourceTrend
GPQA Diamond71.7%#175–85.3%#9986.2%#93–Epoch AI Benchmarking Hub
Humanity's Last Exam–––25.3%#1525.3%#15Scale AI / CAIS
SimpleBench–––56.7%#52–SimpleBench
Arena Hard Prompts–––1448#1161449#115Arena (LMArena)
GPQA Diamond (AA)not in index67.3%#32780.8%#19284.2%#15185.3%#13168.6%#317Artificial Analysis
Humanity's Last Exam (AA)not in index6.0%#43019.6%#23725.4%#20028.5%#1786.6%#406Artificial Analysis
GPQA Diamond (Vals)not in index–––85.6%#45–Vals AI
Kagi LLM Benchmarkeffort not stated––––72.7%#21Kagi LLM Benchmark
GPQA Diamond (HELM)not in indexeffort not stated––––79.1%#2HELM Capabilities
ARC-AGI-16.0%#24944.0%#19656.2%#17965.7%#159–ARC Prize
ARC-AGI-20.0%#2511.9%#2137.5%#1749.9%#168–ARC Prize
CritPtnot in index0.0%#3371.1%#2420.0%#3375.7%#163–Artificial Analysis
CritPt v2 hard subset (AA)–––11.5%#108–Artificial Analysis
Arena Expertnot in index–––1460#1071445#125Arena (LMArena)
Chess Puzzlesnot in index11.0%#12827.0%#7732.0%#6836.0%#53–Epoch AI Benchmarking Hub
Mystery Game Puzzlesnot in index14.0%#10216.0%#8915.0%#9623.0%#55–Epoch AI Benchmarking Hub
LMCAnot in index–––40.0%#113–Epoch AI Benchmarking Hub
DTBenchnot in index–––90.7%#85–Epoch AI Benchmarking Hub
ForecastBenchnot in indexeffort not stated––––65.0%#17ForecastBench

Coding

Benchmarkminimallowmediumhighnot statedSourceTrend
SWE-bench Verified (Epoch)––71.5%#2373.5%#19–Epoch AI Benchmarking Hub
SciCodeeffort not stated––––42.9%#165SciCode
WeirdML–––60.7%#4839.8%#104WeirdML
GSO-Bench–––6.9%#23–GSO-Bench
Arena Coding–––1471#1221462#132Arena (LMArena)
Arena WebDev––1417#95––Arena (LMArena)
LiveCodeBench–––85.9%#27–Vals AI
SWE-bench (Vals)not in index–––69.0%#65–Vals AI
SWE-Bench Pro–––41.8%#10–Scale AI SEAL
Aider Polyglot–81.3%#586.7%#288.0%#1–Aider polyglot leaderboard
SWE-bench Verified (bash only)––65.0%#20––SWE-bench
SWE-bench Verified (any scaffold)not in indexeffort not stated––––74.4%#8SWE-bench
Vibe Code Bench v1.1not in index–––20.1%#91–Vals AI
SWE-Bench Pro (private)not in indexeffort not stated––––14.9%#11Scale AI SEAL
MirrorCodenot in index–––4.4%#45–Epoch AI Benchmarking Hub
ALE-Benchnot in index807.6#83––1162.5#54–Epoch AI Benchmarking Hub
AlgoTunenot in index–––1.7#8–AlgoTune
Terminal-Bench 2.1 (AA)not in index–––35.2%#151–Artificial Analysis
GameBench 2 by reasoning settingnot in index–39.7#13838.0#14646.2#109–DuelLab GameBench 2
Design Arena Websitenot in index1176#102––1174#105–Design Arena
LiveCodeBench (AA)not in index55.8%#12176.3%#4270.3%#6484.5%#1654.3%#127Artificial Analysis

Agents & tools

Benchmarkminimallowmediumhighnot statedSourceTrend
Terminal-Bench––49.6%#42–49.6%#41Terminal-Bench
GDPval––34.8%#6––OpenAI
Remote Labor Indexeffort not stated––––1.7%#16Scale AI / CAIS
Terminal-Bench Hard18.2%#17526.5%#13237.9%#6232.6%#9812.9%#211Artificial Analysis
τ²-Bench Telecom (AA)not in index67.0%#16784.2%#10586.5%#8784.8%#980.0%#398Artificial Analysis
GDPval-AA v2.1not in index–––925#198–Artificial Analysis
τ³-Banking (AA)–––22.1%#92–Artificial Analysis
Poker Agentnot in indexeffort not stated––––1103.2%#2Vals AI
Terminal-Bench 2.0 (Vals)not in index–––37.1%#37–Vals AI
Arena Searchnot in indexeffort not stated––––1189#27Arena (LMArena)
BALROGnot in index32.8%#14––––BALROG
DeepResearch Benchnot in index48.9%#1649.6%#1248.6%#1848.1%#20–DeepResearch Bench
METR Time Horizonsnot in index––69.6%#1069.4%#11–METR
BrowseComp-Plusnot in indexeffort not stated––––57.6%#1BrowseComp-Plus
BrowseCompnot in indexeffort not stated––––20.1%#2OpenAI
DeepSearchQAnot in indexeffort not stated––––59.4%#5Google

Maths

Benchmarkminimallowmediumhighnot statedSourceTrend
FrontierMath Tiers 1–318.3%#10037.2%#68–55.4%#53–Epoch AI Benchmarking Hub
FrontierMath Tier 4–––21.9%#56–Epoch AI Benchmarking Hub
OTIS Mock AIME46.7%#199–87.2%#9591.4%#77–Epoch AI Benchmarking Hub
MATH Level 5––97.9%#298.1%#1–Epoch AI Benchmarking Hub
ProofBench–––18.0%#56–Vals AI
AIME (Vals)–––93.4%#14–Vals AI
MGSM–––92.8%#14–Vals AI
IMO 2025possible contamination–––38.1%#1–MathArena
MathArena Apex–––1.0%#36–MathArena
Omni-MATH (HELM)effort not stated––––64.7%#7HELM Capabilities
MATH 500not in index–––96.0%#3–Vals AI
Arena Maths–––1435#1021411#140Arena (LMArena)
AIME 2025 (AA)not in index31.7%#17283.0%#5391.7%#1894.3%#1248.3%#137Artificial Analysis

Knowledge

Benchmarkminimalmediumhighnot statedSourceTrend
SimpleQA Verified––50.1%#31–Epoch AI Benchmarking Hub
MMLU-Pro––86.5%#40–Vals AI
LegalBench––86.0%#18–Vals AI
CorpFin––61.1%#59–Vals AI
TaxEval––73.4%#47–Vals AI
MedQA––96.3%#4–Vals AI
PRBench Financeeffort not stated–––51.3%#7Scale AI SEAL
PRBench Legaleffort not stated–––49.0%#16Scale AI SEAL
MultiNRCeffort not statedpossible contamination–––52.1%#9Scale AI SEAL
MMLU-Pro (HELM)not in indexeffort not stated–––86.3%#4HELM Capabilities
CaseLaw v2not in index––66.5%#6–Vals AI
MedCodenot in index––49.6%#21–Vals AI
MedScribenot in index––83.7%#43–Vals AI
EBR-benchnot in index––12.7%#24–Epoch AI Benchmarking Hub
Global-MMLU-Lite (AA)not in index84.4%#4390.4%#1290.7%#7–Artificial Analysis

Instruction following

Benchmarkminimallowmediumhighthinkingnot statedSourceTrend
IFBench45.6%#20566.6%#9370.6%#6673.1%#44–45.0%#211Artificial Analysis
MultiChallenge––––63.2%#13–Scale AI SEAL
TutorBencheffort not stated–––––55.3%#4Scale AI SEAL
IFEval (HELM)effort not stated–––––87.5%#15HELM Capabilities
WildBench (HELM)effort not stated–––––85.7%#6HELM Capabilities
Arena Instruction Followingnot in index–––1410#130–1415#121Arena (LMArena)

Human preference

Benchmarkmediumhighnot statedSourceTrend
Arena Text–1435#1061427#117Arena (LMArena)
EQ-Bench Creative Writing v3not in indexeffort not stated––1626#43EQ-Bench
Arena Creative Writingnot in index–1373#1511389#127Arena (LMArena)
Arena Multi-turnnot in index–1421#1291449#90Arena (LMArena)
Arena Longer Queriesnot in index–1413#1471435#118Arena (LMArena)
Lech Mazur Creative Writingnot in index86.0%#1––Lech Mazur

Multimodal

Benchmarkminimallowmediumhighthinkingnot statedSourceTrend
Arena Vision–––1208#89–1232#77Arena (LMArena)
MMMU-Pro62.1%#21373.8%#13374.3%#12474.2%#126––Artificial Analysis
VISTAeffort not stated–––––49.7%#11Scale AI SEAL
MMMU-Pro (Vals)not in index–––81.5%#39––Vals AI
MortgageTaxnot in index–––65.5%#40––Vals AI
VTBnot in index––––18.7%#617.0%#7Scale AI SEAL
GeoBenchnot in index––81.0%#11–––GeoBench
VPCTnot in index––63.2%#666.0%#5––Epoch AI Benchmarking Hub

Long context

Benchmarkmediumhighnot statedSourceTrend
Fiction.LiveBench 120k96.9%#2––Fiction.live
AA-LCR v1.176.0%#16578.2%#13665.0%#270Artificial Analysis

Composite

Benchmarkminimallowmediumhighnot statedSourceTrend
Epoch Capabilities Indexnot in indexeffort not stated––––150#67Epoch AI Benchmarking Hub
AA Intelligence Index v4.3.2not in index11.4#36220.8#23122.9#20223.0#19910.4#386Artificial Analysis
AA Progress Index v1not in index194.9#438254#253264.4#229270.7#200210.7#381Artificial Analysis
HELM Capabilities meannot in indexeffort not stated––––80.7%#4HELM Capabilities

Safety & honesty

Benchmarknot statedSourceTrend
FORTRESSnot in index17.0%#30Scale AI SEAL
MASKnot in indexpossible contamination79.3%#25Scale AI SEAL

GameBench 2, best reasoning setting (High): 46.2. Reference only, not part of the BenchLeader Index.

DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

Where it wins

Benchmarks where this configuration ranks in the top five of every configuration measured.

API pricing: where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

ProviderSpeedFirst tokenInput $/MOutput $/MContextQuantisation
Azure88 tok/s28 s$1.25$10.00400k–
OpenAI78 tok/s7.59 s$1.25$10.00400k–
Azure41 tok/s3.64 s$1.38$11.00400k–

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.125 per 1M. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.0035$0.003229.3 s
Summarize a 30-page report12,000 / 600$0.021$0.01133.4 s
Code edit6,000 / 1,500$0.022$0.01745.5 s
Agentic coding session60,000 / 4,000$0.115$0.0641.3 min
Structured extraction2,000 / 200$0.0045$0.002828.0 s

More from OpenAI

New this month

Compare GPT-5 with…

See also

Embed or cite

Show where GPT-5 ranks on a model card or README. The badge redraws itself with every daily update, carries the data date and links back to this page.

[![GPT-5 on BenchLeader: #51 overall](https://www.benchleader.com/badge/gpt-5.svg)](https://www.benchleader.com/models/gpt-5)

Cite this

The ranking changes daily, so a reference carries the data date. APA style:

BenchLeader. (2026, November 15). GPT-5 on the BenchLeader Index (method v1.4) [Data set]. Retrieved November 15, 2026, from https://www.benchleader.com/models/gpt-5

BibTeX:

@misc{benchleader_gpt_5_20261115,
  author       = {{BenchLeader}},
  title        = {{GPT-5 on the BenchLeader Index (method v1.4)}},
  year         = {2026},
  month        = nov,
  howpublished = {\url{https://www.benchleader.com/models/gpt-5}},
  note         = {Data as of 2026-11-15}
}

Every badge style, the embeddable top-10 widget and how to cite the Index

Questions people ask

What are GPT-5's benchmark scores?
As of 15 Nov 2026, GPT-5 (medium effort) scores 59.0 ±5.4 on the BenchLeader Index, strongest in long context (70) and weakest in reasoning (48). That rests on independent benchmark results only, and every score on this page links to its publisher.
How does GPT-5 rank?
GPT-5 (medium effort) is #113 of 608 ranked configurations on 15 Nov 2026. 258 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Fable 5.2 (max effort) at 72.7.
How much does GPT-5 cost per 1M tokens?
As of 15 Nov 2026, GPT-5 costs $1.25 per 1M input tokens and $10.00 per 1M output tokens, $3.44 blended (three input tokens per output token).
Is GPT-5 better than Mimo v2 Pro?
As of 15 Nov 2026, GPT-5 (medium) and Mimo v2 Pro are neck and neck on quality (59.0 vs 59.1, inside the ±7.2 margin of error) and close on price, so pick on the category that matters to you. The head-to-head page has the benchmark-by-benchmark detail.
When was GPT-5 released?
OpenAI released GPT-5 on 7 Aug 2025. BenchLeader has results for it at 6 reasoning-effort settings.

Data as of 15 Nov 2026. Compare these configurations.

Cite as: BenchLeader, “GPT-5: benchmarks, pricing, speed and rank”, data as of 15 Nov 2026, https://www.benchleader.com/models/gpt-5.