BenchAtlas

Rankings / OpenAI

GPT 5.2 (2025-12-11)

proprietary

released 2025-12-11

BenchAtlas Index

as of 2026-07-14
76.6
xhigh effort
rank #27
12 families · 5 categories · high
73.3
xhigh effort
rank #48
10 families · 4 categories · medium
70.0
medium effort
rank #64
12 families · 5 categories · high

Benchmark evidence

177 results
agentic coding
Release 2026-06-25
tasks_counted=3 · livebench_version=2026-06-25
high effort50.3 %independentLiveBench
τ²-Benchxhigh effort84.8 %independentArtificial Analysis
τ²-Benchmedium effort74.3 %independentArtificial Analysis
τ²-Benchno reasoning46.5 %independentArtificial Analysis
τ²-Bench
subset=banking
no reasoning12.0 %independentArtificial Analysis
coding
Release 2026-06-25
tasks_counted=2 · livebench_version=2026-06-25
high effort76.1 %independentLiveBench
LiveCodeBench v6
implementation=artificial-analysis
medium effort89.4 %independentArtificial Analysis
LiveCodeBench v6
implementation=artificial-analysis
xhigh effort88.9 %independentArtificial Analysis
LiveCodeBench v6
implementation=artificial-analysis
no reasoning66.9 %independentArtificial Analysis
SciCodexhigh effort52.1 %independentArtificial Analysis
SciCodemedium effort46.2 %independentArtificial Analysis
SciCodeno reasoning40.4 %independentArtificial Analysis
data analysis
Release 2026-06-25
tasks_counted=3 · livebench_version=2026-06-25
high effort78.2 %independentLiveBench
external indices
Intelligence Index v4.1xhigh effort42.2 pointsindependentArtificial Analysis
Intelligence Index v4.1medium effort38.0 pointsindependentArtificial Analysis
Intelligence Index v4.1no reasoning26.0 pointsindependentArtificial Analysis
AA Math Indexxhigh effort99.0 pointsindependentArtificial Analysis
AA Math Indexmedium effort96.7 pointsindependentArtificial Analysis
AA Math Indexno reasoning51.0 pointsindependentArtificial Analysis
ECImedium effort153.7 pointsindependentEpoch AI Benchmarking Hub
ECIxhigh effort153.7 pointsindependentEpoch AI Benchmarking Hub
ECIno reasoning153.7 pointsindependentEpoch AI Benchmarking Hub
ECIhigh effort153.7 pointsindependentEpoch AI Benchmarking Hub
ECIlow effort153.7 pointsindependentEpoch AI Benchmarking Hub
ECI153.7 pointsindependentEpoch AI Benchmarking Hub
factuality
MASK
contamination=Potential contamination warning: This model was evaluated after the public release of MASK, allowing model builder access to the prompts and solutions.
86.7 %
84.9–88.4
independentScale Labs
SimpleQA Verifiedxhigh effort38.9 %independentEpoch AI Benchmarking Hub
2025-12-11
SimpleQA Verifiedhigh effort38.2 %independentEpoch AI Benchmarking Hub
2025-12-11
SimpleQA Verifiedmedium effort35.4 %independentEpoch AI Benchmarking Hub
2025-12-11
SimpleQA Verifiedlow effort34.7 %independentEpoch AI Benchmarking Hub
2025-12-11
human preference
Coding (style control)older version
arena=text · category=coding · style_control=true
high effort1489.1
1482.8–1495.3
communityLMArena Leaderboard Dataset
Coding (style control)older version
arena=text · category=coding · style_control=true
1482.0
1476.6–1487.5
communityLMArena Leaderboard Dataset
Expert (style control)older version
arena=text · category=expert · style_control=true
high effort1477.9
1467.6–1488.3
communityLMArena Leaderboard Dataset
French (style control)older version
arena=text · category=french · style_control=true
high effort1477.1
1458.7–1495.5
communityLMArena Leaderboard Dataset
Industry Software And It Services (style control)older version
arena=text · category=industry_software_and_it_services · style_control=true
high effort1471.8
1466.5–1477.1
communityLMArena Leaderboard Dataset
Industry Software And It Services (style control)older version
arena=text · category=industry_software_and_it_services · style_control=true
1471.8
1467.0–1476.6
communityLMArena Leaderboard Dataset
Chinese (style control)older version
arena=text · category=chinese · style_control=true
high effort1471.1
1458.6–1483.7
communityLMArena Leaderboard Dataset
Chinese (style control)older version
arena=text · category=chinese · style_control=true
1469.0
1458.0–1480.0
communityLMArena Leaderboard Dataset
German (style control)older version
arena=text · category=german · style_control=true
high effort1464.0
1441.8–1486.1
communityLMArena Leaderboard Dataset
French (style control)older version
arena=text · category=french · style_control=true
1464.0
1448.9–1479.0
communityLMArena Leaderboard Dataset
Industry Mathematical (style control)older version
arena=text · category=industry_mathematical · style_control=true
high effort1462.4
1449.7–1475.0
communityLMArena Leaderboard Dataset
Expert (style control)older version
arena=text · category=expert · style_control=true
1462.1
1453.8–1470.4
communityLMArena Leaderboard Dataset
Hard Prompts English (style control)older version
arena=text · category=hard_prompts_english · style_control=true
high effort1460.4
1454.5–1466.2
communityLMArena Leaderboard Dataset
Industry Medicine And Healthcare (style control)older version
arena=text · category=industry_medicine_and_healthcare · style_control=true
1460.0
1450.9–1469.2
communityLMArena Leaderboard Dataset
Hard Prompts (style control)older version
arena=text · category=hard_prompts · style_control=true
1459.2
1455.1–1463.4
communityLMArena Leaderboard Dataset
Hard Prompts (style control)older version
arena=text · category=hard_prompts · style_control=true
high effort1459.2
1454.6–1463.7
communityLMArena Leaderboard Dataset
Hard Prompts English (style control)older version
arena=text · category=hard_prompts_english · style_control=true
1458.4
1453.2–1463.6
communityLMArena Leaderboard Dataset
Math (style control)older version
arena=text · category=math · style_control=true
high effort1457.5
1446.5–1468.6
communityLMArena Leaderboard Dataset
Polish (style control)older version
arena=text · category=polish · style_control=true
1455.8
1439.6–1471.9
communityLMArena Leaderboard Dataset
Polish (style control)older version
arena=text · category=polish · style_control=true
high effort1455.6
1437.9–1473.4
communityLMArena Leaderboard Dataset
Industry Medicine And Healthcare (style control)older version
arena=text · category=industry_medicine_and_healthcare · style_control=true
high effort1453.6
1442.6–1464.7
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
1451.5
1445.1–1457.8
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 effort1450.8
1443.5–1458.1
communityLMArena Leaderboard Dataset
Industry Legal And Government (style control)older version
arena=text · category=industry_legal_and_government · style_control=true
1447.5
1438.9–1456.2
communityLMArena Leaderboard Dataset
Industry Legal And Government (style control)older version
arena=text · category=industry_legal_and_government · style_control=true
high effort1447.2
1436.7–1457.7
communityLMArena Leaderboard Dataset
English (style control)older version
arena=text · category=english · style_control=true
high effort1444.4
1439.5–1449.3
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 effort1444.4
1437.6–1451.2
communityLMArena Leaderboard Dataset
German (style control)older version
arena=text · category=german · style_control=true
1444.0
1424.6–1463.5
communityLMArena Leaderboard Dataset
Multi Turn (style control)older version
arena=text · category=multi_turn · style_control=true
1444.0
1438.0–1450.0
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
1443.1
1437.1–1449.1
communityLMArena Leaderboard Dataset
Multi Turn (style control)older version
arena=text · category=multi_turn · style_control=true
high effort1443.0
1435.9–1450.0
communityLMArena Leaderboard Dataset
Longer Query (style control)older version
arena=text · category=longer_query · style_control=true
1440.6
1435.6–1445.7
communityLMArena Leaderboard Dataset
Russian (style control)older version
arena=text · category=russian · style_control=true
1439.3
1431.9–1446.8
communityLMArena Leaderboard Dataset
English (style control)older version
arena=text · category=english · style_control=true
1439.0
1434.6–1443.5
communityLMArena Leaderboard Dataset
Overall (style control)
arena=text · category=overall · style_control=true
high effort1437.4
1433.7–1441.0
communityLMArena Leaderboard Dataset
Industry Mathematical (style control)older version
arena=text · category=industry_mathematical · style_control=true
1436.7
1426.3–1447.2
communityLMArena Leaderboard Dataset
Overall (style control)
arena=text · category=overall · style_control=true
1435.3
1431.9–1438.6
communityLMArena Leaderboard Dataset
Spanish (style control)older version
arena=text · category=spanish · style_control=true
high effort1434.6
1417.1–1452.1
communityLMArena Leaderboard Dataset
Longer Query (style control)older version
arena=text · category=longer_query · style_control=true
high effort1434.2
1428.5–1439.9
communityLMArena Leaderboard Dataset
Exclude Ties (style control)older version
arena=text · category=exclude_ties · style_control=true
high effort1434.1
1429.1–1439.0
communityLMArena Leaderboard Dataset
Math (style control)older version
arena=text · category=math · style_control=true
1432.6
1422.9–1442.3
communityLMArena Leaderboard Dataset
Exclude Ties (style control)older version
arena=text · category=exclude_ties · style_control=true
1432.2
1427.6–1436.7
communityLMArena Leaderboard Dataset
Russian (style control)older version
arena=text · category=russian · style_control=true
high effort1430.3
1421.8–1438.9
communityLMArena Leaderboard Dataset
Instruction Following (style control)older version
arena=text · category=instruction_following · style_control=true
high effort1427.1
1421.4–1432.8
communityLMArena Leaderboard Dataset
Non English (style control)older version
arena=text · category=non_english · style_control=true
1426.2
1422.0–1430.5
communityLMArena Leaderboard Dataset
Non English (style control)older version
arena=text · category=non_english · style_control=true
high effort1425.7
1421.1–1430.3
communityLMArena Leaderboard Dataset
Instruction Following (style control)older version
arena=text · category=instruction_following · style_control=true
1423.1
1418.0–1428.1
communityLMArena Leaderboard Dataset
Spanish (style control)older version
arena=text · category=spanish · style_control=true
1419.1
1404.2–1433.9
communityLMArena Leaderboard Dataset
Japanese (style control)older version
arena=text · category=japanese · style_control=true
1413.4
1387.7–1439.2
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
1411.6
1406.0–1417.2
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 effort1409.0
1402.6–1415.3
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
1397.8
1391.9–1403.8
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 effort1390.4
1383.4–1397.3
communityLMArena Leaderboard Dataset
Creative Writing (style control)older version
arena=text · category=creative_writing · style_control=true
1389.8
1383.2–1396.3
communityLMArena Leaderboard Dataset
Creative Writing (style control)older version
arena=text · category=creative_writing · style_control=true
high effort1388.0
1380.3–1395.6
communityLMArena Leaderboard Dataset
Korean (style control)older version
arena=text · category=korean · style_control=true
1387.1
1367.9–1406.3
communityLMArena Leaderboard Dataset
Korean (style control)older version
arena=text · category=korean · style_control=true
high effort1378.8
1356.0–1401.7
communityLMArena Leaderboard Dataset
knowledge science
GPQA Diamond
implementation=vals-ai
xhigh effort91.7 %independentVals AI
GPQA Diamondxhigh effort91.4 %independentEpoch AI Benchmarking Hub
2025-12-13
GPQA Diamond
implementation=artificial-analysis
xhigh effort90.3 %independentArtificial Analysis
GPQA Diamondhigh effort88.2 %independentEpoch AI Benchmarking Hub
2025-12-11
GPQA Diamondmedium effort87.9 %independentEpoch AI Benchmarking Hub
2025-12-11
GPQA Diamond
implementation=artificial-analysis
medium effort86.4 %independentArtificial Analysis
GPQA Diamondlow effort82.7 %independentEpoch AI Benchmarking Hub
2025-12-11
GPQA Diamond
implementation=artificial-analysis
no reasoning71.2 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
xhigh effort35.4 %independentArtificial Analysis
Humanity's Last Exam
contamination=Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. · implementation=scale
27.8 %
26.0–29.6
independentScale Labs
Humanity's Last Exam
implementation=artificial-analysis
medium effort24.9 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
no reasoning7.3 %independentArtificial Analysis
MMLU-Pro
implementation=artificial-analysis
xhigh effort87.4 %independentArtificial Analysis
MMLU-Pro
implementation=vals-ai
xhigh effort86.2 %independentVals AI
MMLU-Pro
implementation=artificial-analysis
medium effort85.9 %independentArtificial Analysis
MMLU-Pro
implementation=artificial-analysis
no reasoning81.4 %independentArtificial Analysis
language
Release 2026-06-25
tasks_counted=3 · livebench_version=2026-06-25
high effort79.8 %independentLiveBench
long context instruction
AA-LCRxhigh effort72.7 %independentArtificial Analysis
AA-LCRmedium effort63.3 %independentArtificial Analysis
AA-LCRno reasoning38.0 %independentArtificial Analysis
IFBenchxhigh effort75.4 %independentArtificial Analysis
IFBenchmedium effort65.2 %independentArtificial Analysis
IFBenchno reasoning47.4 %independentArtificial Analysis
Release 2026-06-25
tasks_counted=4 · livebench_version=2026-06-25
high effort61.8 %independentLiveBench
multimodal
MMMU
implementation=vals-ai
xhigh effort86.7 %independentVals AI
VISTA46.6 %
44.4–48.9
independentScale Labs
professional
CaseLawxhigh effort66.0 %independentVals AI
CorpFinxhigh effort65.9 %independentVals AI
LegalBenchxhigh effort82.8 %independentVals AI
MedQAxhigh effort94.1 %independentVals AI
TaxEvalxhigh effort75.8 %independentVals AI
reasoning math
AIME
year=2025 · implementation=artificial-analysis
xhigh effort99.0 %independentArtificial Analysis
AIME
implementation=vals-ai
xhigh effort96.9 %independentVals AI
AIME
year=2025 · implementation=artificial-analysis
medium effort96.7 %independentArtificial Analysis
AIME
year=2025 · implementation=artificial-analysis
no reasoning51.0 %independentArtificial Analysis
ARC-AGI-1older version
split=public_eval · model_type=CoT
max effort95.0 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
high effort90.3 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
max effort86.2 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
medium effort80.6 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
high effort78.7 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
medium effort72.7 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
low effort65.9 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
max effort59.8 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
low effort55.7 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
max effort52.9 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
high effort43.3 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
high effort39.9 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
medium effort27.6 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
medium effort26.7 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=Base LLM
no reasoning16.5 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=Base LLM
no reasoning12.3 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
low effort9.7 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
low effort8.3 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
max effort2.1 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
max effort1.9 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
high effort1.5 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
high effort1.4 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
max effort1.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=Base LLM
no reasoning0.8 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
medium effort0.8 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
medium effort0.8 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
max effort0.7 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
high effort0.5 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
high effort0.4 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
medium effort0.3 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
medium effort0.3 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
low effort0.3 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
low effort0.3 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
low effort0.2 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
low effort0.1 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=Base LLM
no reasoning0.1 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=Base LLM
no reasoning0.1 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=Base LLM
no reasoning0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=Base LLM
no reasoning0.0 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=Base LLM
no reasoning0.0 %independentARC Prize Leaderboard
EnigmaEval
contamination=Potential contamination warning: This model was evaluated after the public release of EnigmaEval, allowing model builder access to the prompts and solutions.
10.4 %
8.7–12.1
independentScale Labs
FrontierMathxhigh effort40.7 %independentEpoch AI Benchmarking Hub
2025-12-13
FrontierMathhigh effort40.3 %independentEpoch AI Benchmarking Hub
2025-12-11
FrontierMathmedium effort36.9 %independentEpoch AI Benchmarking Hub
2025-12-11
FrontierMathlow effort26.6 %independentEpoch AI Benchmarking Hub
2025-12-11
FrontierMath Tier 4older versionxhigh effort18.8 %independentEpoch AI Benchmarking Hub
2025-12-14
FrontierMath Tier 4older versionhigh effort18.8 %independentEpoch AI Benchmarking Hub
2025-12-11
FrontierMath Tier 4older versionmedium effort16.7 %independentEpoch AI Benchmarking Hub
2025-12-11
FrontierMath Tier 4older versionlow effort6.3 %independentEpoch AI Benchmarking Hub
2025-12-11
Release 2026-06-25
tasks_counted=4 · livebench_version=2026-06-25
high effort93.2 %independentLiveBench
Release 2026-06-25
tasks_counted=4 · livebench_version=2026-06-25
high effort83.2 %independentLiveBench
OTIS Mock AIME 2024–2025high effort96.1 %independentEpoch AI Benchmarking Hub
2025-12-11
OTIS Mock AIME 2024–2025xhigh effort96.1 %independentEpoch AI Benchmarking Hub
2025-12-13
OTIS Mock AIME 2024–2025medium effort93.9 %independentEpoch AI Benchmarking Hub
2025-12-11
OTIS Mock AIME 2024–2025low effort78.9 %independentEpoch AI Benchmarking Hub
2025-12-11

Agent + model results

systems, not bare-model scores
agent + model Epoch Inspect harness + GPT 5.2 (2025-12-11)SWE-bench Verified73.8 %independentEpoch AI Benchmarking Hub
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench bash-only72.8 %unverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Verified72.8 %unverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench bash-only71.8 %unverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Verified71.8 %unverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Verified69.0 %unverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench bash-only69.0 %unverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Multilingual66.7 %communitySWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Multilingual0.5 usd_per_taskcommunitySWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench bash-only0.5 usd_per_taskunverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Verified0.5 usd_per_taskunverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench bash-only0.5 usd_per_taskunverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Verified0.5 usd_per_taskunverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench Verified0.3 usd_per_taskunverifiedSWE-bench Leaderboard
agent + model mini-SWE-agent + GPT 5.2 (2025-12-11)SWE-bench bash-only0.3 usd_per_taskunverifiedSWE-bench Leaderboard
agent + model Droid + GPT 5.2 (2025-12-11)Terminal-Bench 2.064.9 %unverifiedTerminal-Bench Leaderboard
agent + model Codex CLI + GPT 5.2 (2025-12-11)Terminal-Bench 2.062.9 %communityTerminal-Bench Leaderboard
agent + model mux + GPT 5.2 (2025-12-11)Terminal-Bench 2.060.7 %unverifiedTerminal-Bench Leaderboard
agent + model terminus-2 + GPT 5.2 (2025-12-11)Terminal-Bench 2.054.0 %communityTerminal-Bench Leaderboard
agent + model Artificial Analysis harness + GPT 5.2 (2025-12-11)Terminal-Bench Hard47.0 %independentArtificial Analysis
agent + model Artificial Analysis harness + GPT 5.2 (2025-12-11)Terminal-Bench Hard43.2 %independentArtificial Analysis
agent + model Artificial Analysis harness + GPT 5.2 (2025-12-11)Terminal-Bench Hard31.8 %independentArtificial Analysis

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

Related variants

GPT 4.5 (2025-02-27)GPT 4o (2024-11-20)GPT 4.1 (2025-04-14)GPT 5 (2025-08-07)GPT 5.1 (2025-11-13)GPT 5.5 (2026-04-22)GPT 5.4 (2026-03-05)GPT 4 0314