BenchAtlas

Rankings / OpenAI

GPT 5 Nano (2025-08-07)

proprietary

released 2025-08-07

BenchAtlas Index

as of 2026-07-14
31.3
high effort
rank #347
14 families · 5 categories · high
30.9
high effort
rank #348
10 families · 5 categories · high
28.2
medium effort
rank #369
14 families · 5 categories · high

Benchmark evidence

107 results
agentic coding
τ²-Benchhigh effort36.5 %independentArtificial Analysis
τ²-Benchmedium effort30.4 %independentArtificial Analysis
τ²-Benchminimal effort25.7 %independentArtificial Analysis
coding
LiveCodeBench v6
implementation=artificial-analysis
high effort78.9 %independentArtificial Analysis
LiveCodeBench v6
implementation=artificial-analysis
medium effort76.3 %independentArtificial Analysis
LiveCodeBench v6
implementation=artificial-analysis
minimal effort47.0 %independentArtificial Analysis
SciCodehigh effort36.6 %independentArtificial Analysis
SciCodemedium effort33.8 %independentArtificial Analysis
SciCodeminimal effort29.1 %independentArtificial Analysis
external indices
Intelligence Index v4.1high effort19.9 pointsindependentArtificial Analysis
Intelligence Index v4.1medium effort19.0 pointsindependentArtificial Analysis
Intelligence Index v4.1minimal effort8.0 pointsindependentArtificial Analysis
AA Math Indexhigh effort83.7 pointsindependentArtificial Analysis
AA Math Indexmedium effort78.3 pointsindependentArtificial Analysis
AA Math Indexminimal effort27.3 pointsindependentArtificial Analysis
ECIlow effort140.8 pointsindependentEpoch AI Benchmarking Hub
ECI140.8 pointsindependentEpoch AI Benchmarking Hub
ECIminimal effort140.8 pointsindependentEpoch AI Benchmarking Hub
ECImedium effort140.8 pointsindependentEpoch AI Benchmarking Hub
ECIhigh effort140.8 pointsindependentEpoch AI Benchmarking Hub
factuality
SimpleQA Verifiedhigh effort12.2 %independentEpoch AI Benchmarking Hub
2025-12-09
human preference
Coding (style control)older version
arena=text · category=coding · style_control=true
high effort1382.8
1368.2–1397.3
communityLMArena Leaderboard Dataset
Industry Software And It Services (style control)older version
arena=text · category=industry_software_and_it_services · style_control=true
high effort1371.8
1360.6–1382.9
communityLMArena Leaderboard Dataset
Hard Prompts English (style control)older version
arena=text · category=hard_prompts_english · style_control=true
high effort1367.5
1354.2–1380.9
communityLMArena Leaderboard Dataset
Hard Prompts (style control)older version
arena=text · category=hard_prompts · style_control=true
high effort1354.2
1344.4–1363.9
communityLMArena Leaderboard Dataset
Industry Life And Physical And Social Science (style control)older version
arena=text · category=industry_life_and_physical_and_social_science · style_control=true
high effort1345.8
1329.3–1362.3
communityLMArena Leaderboard Dataset
English (style control)older version
arena=text · category=english · style_control=true
high effort1344.0
1334.6–1353.4
communityLMArena Leaderboard Dataset
Polish (style control)older version
arena=text · category=polish · style_control=true
high effort1339.2
1313.4–1365.0
communityLMArena Leaderboard Dataset
Overall (style control)
arena=text · category=overall · style_control=true
high effort1337.2
1330.3–1344.1
communityLMArena Leaderboard Dataset
Industry Business And Management And Financial Operations (style control)older version
arena=text · category=industry_business_and_management_and_financial_operations · style_control=true
high effort1334.3
1318.8–1349.8
communityLMArena Leaderboard Dataset
Longer Query (style control)older version
arena=text · category=longer_query · style_control=true
high effort1329.3
1314.6–1344.0
communityLMArena Leaderboard Dataset
Non English (style control)older version
arena=text · category=non_english · style_control=true
high effort1328.1
1318.9–1337.3
communityLMArena Leaderboard Dataset
Instruction Following (style control)older version
arena=text · category=instruction_following · style_control=true
high effort1325.6
1312.5–1338.8
communityLMArena Leaderboard Dataset
Multi Turn (style control)older version
arena=text · category=multi_turn · style_control=true
high effort1322.3
1306.7–1337.9
communityLMArena Leaderboard Dataset
Exclude Ties (style control)older version
arena=text · category=exclude_ties · style_control=true
high effort1292.3
1282.7–1301.9
communityLMArena Leaderboard Dataset
Industry Writing And Literature And Language (style control)older version
arena=text · category=industry_writing_and_literature_and_language · style_control=true
high effort1290.1
1275.8–1304.3
communityLMArena Leaderboard Dataset
Industry Entertainment And Sports And Media (style control)older version
arena=text · category=industry_entertainment_and_sports_and_media · style_control=true
high effort1277.2
1261.0–1293.4
communityLMArena Leaderboard Dataset
Creative Writing (style control)older version
arena=text · category=creative_writing · style_control=true
high effort1248.2
1228.1–1268.3
communityLMArena Leaderboard Dataset
knowledge science
GPQA Diamondhigh effort69.4 %independentEpoch AI Benchmarking Hub
2025-10-30
GPQA Diamond
implementation=artificial-analysis
high effort67.6 %independentArtificial Analysis
GPQA Diamondmedium effort67.4 %independentEpoch AI Benchmarking Hub
2025-08-07
GPQA Diamond
implementation=artificial-analysis
medium effort67.0 %independentArtificial Analysis
GPQA Diamond
implementation=vals-ai
high effort63.4 %independentVals AI
GPQA Diamond
implementation=artificial-analysis
minimal effort42.8 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
high effort8.2 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
medium effort7.6 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
minimal effort4.1 %independentArtificial Analysis
MMLU-Pro
implementation=artificial-analysis
high effort78.0 %independentArtificial Analysis
MMLU-Pro
implementation=artificial-analysis
medium effort77.2 %independentArtificial Analysis
MMLU-Pro
implementation=vals-ai
high effort76.1 %independentVals AI
MMLU-Pro
implementation=artificial-analysis
minimal effort55.6 %independentArtificial Analysis
long context instruction
AA-LCRhigh effort41.7 %independentArtificial Analysis
AA-LCRmedium effort40.0 %independentArtificial Analysis
AA-LCRminimal effort20.0 %independentArtificial Analysis
IFBenchhigh effort67.5 %independentArtificial Analysis
IFBenchmedium effort65.9 %independentArtificial Analysis
IFBenchminimal effort32.5 %independentArtificial Analysis
multimodal
MMMU
implementation=vals-ai
high effort70.9 %independentVals AI
professional
CaseLawhigh effort52.6 %independentVals AI
LegalBenchhigh effort50.1 %independentVals AI
MedQAhigh effort93.3 %independentVals AI
TaxEvalhigh effort67.4 %independentVals AI
reasoning math
AIME
year=2025 · implementation=artificial-analysis
high effort83.7 %independentArtificial Analysis
AIME
implementation=vals-ai
high effort81.2 %independentVals AI
AIME
year=2025 · implementation=artificial-analysis
medium effort78.3 %independentArtificial Analysis
AIME
year=2025 · implementation=artificial-analysis
minimal effort27.3 %independentArtificial Analysis
ARC-AGI-1older version
split=public_eval · model_type=CoT
high effort29.7 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
medium effort20.8 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
medium effort20.7 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
high effort16.7 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
low effort11.8 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
low effort4.0 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
minimal effort2.8 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
high effort2.6 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
minimal effort1.5 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
medium effort0.9 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
high effort0.3 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
high effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
high effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
high effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
high effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
medium effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
medium effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
medium effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
medium effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
low effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
low effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
low effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
low effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
minimal effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
minimal effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
minimal effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
minimal effort0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
low effort0.0 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
low effort0.0 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
medium effort0.0 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
minimal effort0.0 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
minimal effort0.0 %independentARC Prize Leaderboard
FrontierMathhigh effort8.3 %independentEpoch AI Benchmarking Hub
2025-10-30
FrontierMathmedium effort7.2 %independentEpoch AI Benchmarking Hub
2025-08-07
FrontierMath Tier 4older versionmedium effort2.1 %independentEpoch AI Benchmarking Hub
2025-08-07
FrontierMath Tier 4older versionhigh effort0.0 %independentEpoch AI Benchmarking Hub
2025-10-30
MATH-500
implementation=vals-ai
high effort93.8 %independentVals AI
MATH Level 5medium effort95.2 %independentEpoch AI Benchmarking Hub
2025-08-20
MATH Level 5high effort94.9 %independentEpoch AI Benchmarking Hub
2025-08-20
OTIS Mock AIME 2024–2025high effort81.1 %independentEpoch AI Benchmarking Hub
2025-10-31
OTIS Mock AIME 2024–2025medium effort74.2 %independentEpoch AI Benchmarking Hub
2025-08-07

Agent + model results

systems, not bare-model scores
agent + model mini-SWE-agent + GPT 5 Nano (2025-08-07)SWE-bench Verified34.8 %communitySWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5 Nano (2025-08-07)SWE-bench bash-only34.8 %communitySWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5 Nano (2025-08-07)SWE-bench bash-only0.0 usd_per_taskcommunitySWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5 Nano (2025-08-07)SWE-bench Verified0.0 usd_per_taskcommunitySWE-bench Leaderboard
agent + model spoox-m + GPT 5 Nano (2025-08-07)Terminal-Bench 2.021.8 %unverifiedTerminal-Bench Leaderboard
agent + model Artificial Analysis harness + GPT 5 Nano (2025-08-07)Terminal-Bench Hard17.4 %independentArtificial Analysis
agent + model terminus-2 + GPT 5 Nano (2025-08-07)Terminal-Bench 1.012.2 %communityTerminal-Bench Leaderboard
agent + model Artificial Analysis harness + GPT 5 Nano (2025-08-07)Terminal-Bench Hard12.1 %independentArtificial Analysis
agent + model Codex CLI + GPT 5 Nano (2025-08-07)Terminal-Bench 2.011.5 %communityTerminal-Bench Leaderboard
agent + model OpenHands + GPT 5 Nano (2025-08-07)Terminal-Bench 2.09.9 %communityTerminal-Bench Leaderboard
agent + model terminus-2 + GPT 5 Nano (2025-08-07)Terminal-Bench 2.07.9 %communityTerminal-Bench Leaderboard
agent + model mini-SWE-agent + GPT 5 Nano (2025-08-07)Terminal-Bench 2.07.0 %communityTerminal-Bench Leaderboard
agent + model Artificial Analysis harness + GPT 5 Nano (2025-08-07)Terminal-Bench Hard6.8 %independentArtificial Analysis

These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.