Rankings / OpenAI

GPT 5.6 Terra

released 2026-07-09

BenchAtlas Index

as of 2026-07-14

max effort

rank #3

14 families · 5 categories · high

xhigh effort

rank #12

7 families · 4 categories · medium

high effort

rank #20

8 families · 5 categories · high

Benchmark evidence

127 results

agentic coding

Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25	max effort	68.0 %	independent	LiveBench
Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25	xhigh effort	53.3 %	independent	LiveBench
τ²-Bench	max effort	86.3 %	independent	Artificial Analysis
τ²-Bench	xhigh effort	80.4 %	independent	Artificial Analysis
τ²-Bench	high effort	78.4 %	independent	Artificial Analysis
τ²-Bench	medium effort	72.8 %	independent	Artificial Analysis
τ²-Bench	low effort	60.5 %	independent	Artificial Analysis
τ²-Bench subset=banking	max effort	31.8 %	independent	Artificial Analysis
τ²-Bench subset=banking	xhigh effort	24.3 %	independent	Artificial Analysis
τ²-Bench subset=banking	high effort	22.3 %	independent	Artificial Analysis
τ²-Bench subset=banking	medium effort	19.4 %	independent	Artificial Analysis
τ²-Bench subset=banking	low effort	16.1 %	independent	Artificial Analysis
τ²-Bench subset=banking	no reasoning	13.4 %	independent	Artificial Analysis

coding

Release 2026-06-25 tasks_counted=2 · livebench_version=2026-06-25	max effort	78.3 %	independent	LiveBench
Release 2026-06-25 tasks_counted=2 · livebench_version=2026-06-25	xhigh effort	75.4 %	independent	LiveBench
SciCode	max effort	53.9 %	independent	Artificial Analysis
SciCode	xhigh effort	51.6 %	independent	Artificial Analysis
SciCode	high effort	50.1 %	independent	Artificial Analysis
SciCode	medium effort	49.7 %	independent	Artificial Analysis
SciCode	low effort	49.2 %	independent	Artificial Analysis
SciCode	no reasoning	44.6 %	independent	Artificial Analysis

data analysis

Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25	max effort	79.3 %	independent	LiveBench
Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25	xhigh effort	77.4 %	independent	LiveBench

external indices

AA Coding Index	max effort	76.7 points	independent	Artificial Analysis
AA Coding Index	xhigh effort	70.6 points	independent	Artificial Analysis
AA Coding Index	high effort	67.1 points	independent	Artificial Analysis
AA Coding Index	medium effort	64.7 points	independent	Artificial Analysis
AA Coding Index	low effort	58.1 points	independent	Artificial Analysis
AA Coding Index	no reasoning	52.3 points	independent	Artificial Analysis
Intelligence Index v4.1	max effort	55.0 points	independent	Artificial Analysis
Intelligence Index v4.1	xhigh effort	51.6 points	independent	Artificial Analysis
Intelligence Index v4.1	high effort	49.0 points	independent	Artificial Analysis
Intelligence Index v4.1	medium effort	45.6 points	independent	Artificial Analysis
Intelligence Index v4.1	low effort	40.5 points	independent	Artificial Analysis
Intelligence Index v4.1	no reasoning	34.0 points	independent	Artificial Analysis
ECI		158.3 points	independent	Epoch AI Benchmarking Hub
ECI	max effort	158.3 points	independent	Epoch AI Benchmarking Hub
ECI	xhigh effort	158.3 points	independent	Epoch AI Benchmarking Hub
ECI	high effort	158.3 points	independent	Epoch AI Benchmarking Hub
ECI	medium effort	158.3 points	independent	Epoch AI Benchmarking Hub
ECI	low effort	158.3 points	independent	Epoch AI Benchmarking Hub
Vals Index	xhigh effort	65.1 points	independent	Vals AI

factuality

SimpleQA Verified

max effort

43.1 %

independent

Epoch AI Benchmarking Hub

2026-07-09

knowledge science

GPQA Diamond	max effort	93.3 %	independent	Epoch AI Benchmarking Hub 2026-07-09
GPQA Diamond implementation=artificial-analysis	max effort	92.5 %	independent	Artificial Analysis
GPQA Diamond implementation=vals-ai	xhigh effort	90.9 %	independent	Vals AI
GPQA Diamond implementation=artificial-analysis	xhigh effort	90.8 %	independent	Artificial Analysis
GPQA Diamond implementation=artificial-analysis	high effort	89.6 %	independent	Artificial Analysis
GPQA Diamond implementation=artificial-analysis	medium effort	87.2 %	independent	Artificial Analysis
GPQA Diamond implementation=artificial-analysis	low effort	84.3 %	independent	Artificial Analysis
GPQA Diamond implementation=artificial-analysis	no reasoning	74.6 %	independent	Artificial Analysis
Humanity's Last Exam implementation=artificial-analysis	max effort	41.8 %	independent	Artificial Analysis
Humanity's Last Exam implementation=artificial-analysis	xhigh effort	40.0 %	independent	Artificial Analysis
Humanity's Last Exam implementation=artificial-analysis	high effort	36.7 %	independent	Artificial Analysis
Humanity's Last Exam implementation=artificial-analysis	medium effort	31.6 %	independent	Artificial Analysis
Humanity's Last Exam implementation=artificial-analysis	low effort	27.4 %	independent	Artificial Analysis
Humanity's Last Exam implementation=artificial-analysis	no reasoning	11.0 %	independent	Artificial Analysis
MMLU-Pro implementation=vals-ai	xhigh effort	86.7 %	independent	Vals AI

language

Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25	max effort	82.9 %	independent	LiveBench
Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25	xhigh effort	79.9 %	independent	LiveBench

long context instruction

AA-LCR	max effort	74.0 %	independent	Artificial Analysis
AA-LCR	high effort	72.3 %	independent	Artificial Analysis
AA-LCR	xhigh effort	71.3 %	independent	Artificial Analysis
AA-LCR	medium effort	68.0 %	independent	Artificial Analysis
AA-LCR	low effort	64.3 %	independent	Artificial Analysis
AA-LCR	no reasoning	50.0 %	independent	Artificial Analysis
IFBench	max effort	71.2 %	independent	Artificial Analysis
IFBench	xhigh effort	66.3 %	independent	Artificial Analysis
IFBench	high effort	64.4 %	independent	Artificial Analysis
IFBench	medium effort	62.2 %	independent	Artificial Analysis
IFBench	low effort	59.7 %	independent	Artificial Analysis
Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25	max effort	64.6 %	independent	LiveBench
Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25	xhigh effort	59.4 %	independent	LiveBench

multimodal

MMMU

implementation=vals-ai

xhigh effort

86.5 %

independent

Vals AI

professional

CorpFin	xhigh effort	65.3 %	independent	Vals AI
LegalBench	xhigh effort	85.1 %	independent	Vals AI
TaxEval	xhigh effort	76.2 %	independent	Vals AI

reasoning math

ARC-AGI-1older version split=public_eval · model_type=CoT	max effort	98.3 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	max effort	96.5 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	high effort	95.8 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	xhigh effort	95.6 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	xhigh effort	94.0 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	high effort	92.0 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	max effort	91.3 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	medium effort	84.8 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	max effort	83.9 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	xhigh effort	82.6 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	medium effort	77.0 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	xhigh effort	74.2 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	high effort	73.9 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	low effort	70.8 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	high effort	67.1 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	low effort	60.2 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	medium effort	38.9 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	medium effort	37.5 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	low effort	18.8 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	low effort	14.9 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	max effort	1.2 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	max effort	1.1 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-3older version split=semi_private · model_type=CoT	max effort	0.8 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	xhigh effort	0.7 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	xhigh effort	0.7 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-3older version split=semi_private · model_type=CoT	xhigh effort	0.7 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	high effort	0.6 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	max effort	0.6 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	high effort	0.6 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-3older version split=semi_private · model_type=CoT	high effort	0.5 %	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	max effort	0.3 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	medium effort	0.3 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	medium effort	0.3 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	xhigh effort	0.3 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	xhigh effort	0.2 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	high effort	0.2 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-2 split=public_eval · model_type=CoT	low effort	0.2 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	low effort	0.2 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	high effort	0.1 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	medium effort	0.1 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	medium effort	0.1 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	low effort	0.1 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=public_eval · model_type=CoT	low effort	0.1 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-3older version split=semi_private · model_type=CoT	medium effort	0.1 %	independent	ARC Prize Leaderboard
ARC-AGI-3older version split=semi_private · model_type=CoT	low effort	0.0 %	independent	ARC Prize Leaderboard
Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25	max effort	94.9 %	independent	LiveBench
Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25	xhigh effort	89.5 %	independent	LiveBench
Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25	max effort	90.6 %	independent	LiveBench
Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25	xhigh effort	84.9 %	independent	LiveBench
OTIS Mock AIME 2024–2025	max effort	99.7 %	independent	Epoch AI Benchmarking Hub 2026-07-09

Agent + model results

systems, not bare-model scores

agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench 2.1	88.0 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench 2.1	80.2 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench 2.1	75.7 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench 2.1	72.3 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench Hard	62.9 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench 2.1	62.5 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench Hard	57.6 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench Hard	57.6 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench 2.1	56.2 %	independent	Artificial Analysis
agent + model Artificial Analysis harness + GPT 5.6 Terra	Terminal-Bench Hard	43.9 %	independent	Artificial Analysis

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