RankingEffort benchmarks

GPT-5

Compare models

85 measured results · 5 effort settings · 23 benchmark/harness combinations

Model evidence and sources →

Change model

Model

Current configurationSources

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare withModel and configuration

Capability profile

One configuration, seven capabilities

Comparison score · 0–100

AGPT-5 · High4 of 7 capabilities supported

Gaps are unknown, not zero. Hollow points are preliminary.

Agentic
A21.9Supported
Hard reasoning
A21.0Supported
Coding
A15.5Supported
Human pref
Unknown
Knowledge
A31.4Supported
Multimodal
A20.5Preliminary
Long context
A42.1Preliminary

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Capability score evidence

GPT-5 · High: 11.7 overall comparison score. Meets the overall evidence requirements.

6 areas measured. The available task mix affects this estimate; missing areas remain unknown.

Which sources carry the weight?

Artificial Analysis · 46.2% direct comparison weight · independent of the model provider
ARC Prize · 40.9% direct comparison weight · independent of the model provider
SWE-rebench · 12.9% direct comparison weight · independent of the model provider

Which benchmark families carry the weight?

ARC-AGI · 40.9% direct comparison weight
swe-rebench · 12.9% direct comparison weight
GDPval · 3.5% direct comparison weight
Tau Bench · 3.3% direct comparison weight
Gmmlu · 3.3% direct comparison weight
Other families (12) · 36.0% direct comparison weight

These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.

How sensitive is this score?

The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-10-23.

Both ranks use the same surviving configurations for each removal. The cohort starts with 444 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.

What changes when a source is removed?
Removed source	Rank before → after	Rank movement	Cohort remaining / lost support	Score after (change)
Artificial Analysis	142 → 149	Down 7	203 remaining / 241 lost	8.1 (-3.6 points)
ARC Prize	180 → 171	Up 9	444 remaining / 0 lost	16.3 (+4.6 points)
Datacurve	180 → 181	Down 1	444 remaining / 0 lost	9.3 (-2.5 points)
LiveCodeBench	180 → 179	Up 1	444 remaining / 0 lost	11.8 (+0.0 points)
Scale	180 → 180	No movement	444 remaining / 0 lost	13.3 (+1.6 points)
Terminal-Bench	180 → 180	No movement	444 remaining / 0 lost	11.6 (-0.1 points)
Cognition	180 → 182	Down 2	444 remaining / 0 lost	11.2 (-0.5 points)
LiveBench	179 → 177	Up 2	443 remaining / 1 lost	11.9 (+0.2 points)
SWE-rebench	180 → 178	Up 2	444 remaining / 0 lost	12.8 (+1.0 points)
Epoch AI	176 → 181	Down 5	431 remaining / 13 lost	11.8 (+0.0 points)
Vals AI	180 → 180	No movement	443 remaining / 1 lost	11.1 (-0.6 points)
Provider reports · Anthropic	180 → 180	No movement	444 remaining / 0 lost	11.7 (-0.0 points)
Provider reports · Moonshot AI	180 → 180	No movement	444 remaining / 0 lost	11.4 (-0.3 points)
Provider reports · Meta	180 → 180	No movement	444 remaining / 0 lost	11.7 (-0.0 points)
Provider reports · Mistral	180 → 180	No movement	444 remaining / 0 lost	11.7 (+0.0 points)
Provider reports · OpenAI	180 → 180	No movement	444 remaining / 0 lost	11.8 (+0.1 points)
All provider reports	180 → 180	No movement	444 remaining / 0 lost	11.5 (-0.2 points)

Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.

Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit

Every benchmark family

Removed evidence	Recomputed score	Change
arc-agi	15.8	+4.1
deepswe	9.3	-2.5
swe-rebench	12.8	+1.0
hle	12.7	+0.9
mmmu-pro	12.6	+0.9
omniscience	11.0	-0.7
epoch-game-puzzles	11.2	-0.6
arc-agi-3	11.2	-0.5
frontiercode	11.2	-0.5
simpleqa	12.2	+0.5
mmlu-pro	12.1	+0.4
aa-lcr	11.4	-0.3
terminal-bench	12.0	+0.3
ifbench	11.4	-0.3
livebench-instructions	12.0	+0.2
gmmlu	11.5	-0.2
multi-swe-bench	11.5	-0.2
gdpval	11.9	+0.2
livebench-language	11.9	+0.2
livecodebench	11.9	+0.2
gpqa	11.6	-0.1
harvey	11.6	-0.1
gdp-pdf	11.9	+0.1
enigma-eval	11.6	-0.1
swe-atlas-qna	11.6	-0.1
apex-agents	11.8	+0.1
livebench-coding	11.6	-0.1
terminal-bench-science	11.8	+0.1
aime	11.6	-0.1
critpt	11.6	-0.1
vibe-code	11.7	-0.1
livebench-reasoning	11.7	-0.1
analyst-agent	11.8	+0.1
automationbench	11.7	-0.1
itbench	11.7	-0.1
briefcase	11.7	-0.1
frontiermath	11.8	+0.1
swe-atlas-test-writing	11.7	-0.0
math500	11.8	+0.0
vals-code-migration	11.7	-0.0
vals-finance-agent	11.7	-0.0
vals-legal-research	11.7	-0.0
vals-tax-agent	11.7	-0.0
vals-excel-modeling	11.7	-0.0
livebench-math	11.7	+0.0
livebench-data	11.7	-0.0
swe-atlas-refactoring	11.7	-0.0
tau-bench	11.7	-0.0
enterprise-ops	11.7	-0.0
swe-bench-pro	11.7	+0.0

What evidence is missing from the fit?

Across all collected settings for this model: 103 observations, 81 contributing, 0 matched without graph weight and 22 excluded. Counts do not establish rank eligibility.

4 · Benchmark family or source protocol has not been reviewed
12 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately
6 · Supporting evidence outside the reviewed capability core

Inspect every observation and exclusion →

Benchmark results

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

Benchmark / source	Minimal	Low	Medium	High	Non-reasoning
AIME 2025Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	31.67% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	83% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	91.67% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	94.33% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	48.33% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	76% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	78.2% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	65% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally. CritPt is under review by its authors following external feedback · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	0% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	1.14% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	0% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	5.71% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	Not reported	Not reported	917.24 Elo Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
Global-MMLU-LiteArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	84.42% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	Not reported	90.39% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	90.67% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
GPQA DiamondArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	67.27% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	80.81% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	84.18% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	85.35% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	68.59% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	6.02% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	19.56% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	25.39% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	28.5% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	6.63% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
IFBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	45.58% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	66.6% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	70.61% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	73.06% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	45.03% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
LiveCodeBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	55.77% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	76.3% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	70.26% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	84.55% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	54.29% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
MATH-500Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	86.13% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	98.73% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	99.13% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	99.4% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
MMLU-ProArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	80.63% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	85.98% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	86.74% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	87.07% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	82.04% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
MMMU-ProArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	62.08% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	73.76% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	74.34% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	74.22% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis Unit: index. Source collected 2026-10-18.	-33.8 Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	-10.78 Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	-10.87 Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	-8.73 Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader Benchmark details AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	29.63% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	38.12% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	39.48% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	40.32% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
𝜏²-Bench TelecomArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	66.96% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	84.21% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	86.55% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	84.8% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	0% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
𝜏³-BankingArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Sierra Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	22.06% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
Terminal-Bench HardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	18.18% Reported settings GPT-5 (Minimal); Model source: https://artificialanalysis.ai/models/gpt-5-minimal	26.52% Reported settings GPT-5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-low	37.88% Reported settings GPT-5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-medium	32.58% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	12.88% Reported settings GPT-5 (ChatGPT); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/gpt-5-chatgpt; Mode from evaluator isReasoning metadata; effort budget not specified
Terminal-Bench v2.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by the Laude Institute and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	35.21% Reported settings GPT-5 (High); Model source: https://artificialanalysis.ai/models/gpt-5	Not reported
ARC-AGI-2ARC Prize · v2 Semi-Private Benchmark details Original public board. Configuration and task coverage may differ; inspect source. Unit: percent. Source collected 2026-10-18.	0% Reported settings GPT-5 (Minimal); Source model ID: gpt-5-2025-08-07-minimal; 120 tasks	1.94% Reported settings GPT-5 (Low); Source model ID: gpt-5-2025-08-07-low; 120 tasks	7.49% Reported settings GPT-5 (Medium); Source model ID: gpt-5-2025-08-07-medium; 120 tasks	9.86% Reported settings GPT-5 (High); Source model ID: gpt-5-2025-08-07-high; 120 tasks	Not reported
SWE-rebench · 2025-08 · toolsSWE-rebench · SWE-rebench standardized ReAct tools Benchmark details Average resolved rate across repeated runs, not best-of-N success. Disjoint complete calendar months within each model’s tested range; standardized harness and all-language task set. Unit: percent. Source collected 2026-10-18.	33.46% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-08-tools; gpt-5-2025-08-07-minimal; checkpoint=gpt-5	Not reported	45.38% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-08-tools; gpt-5-2025-08-07-medium; checkpoint=gpt-5	46.54% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-08-tools; gpt-5-2025-08-07-high; checkpoint=gpt-5	Not reported
SWE-rebench · 2025-09 · toolsSWE-rebench · SWE-rebench standardized ReAct tools Benchmark details Average resolved rate across repeated runs, not best-of-N success. Disjoint complete calendar months within each model’s tested range; standardized harness and all-language task set. Unit: percent. Source collected 2026-10-18.	30.61% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-09-tools; gpt-5-2025-08-07-minimal; checkpoint=gpt-5	Not reported	38.78% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-09-tools; gpt-5-2025-08-07-medium; checkpoint=gpt-5	36.33% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-09-tools; gpt-5-2025-08-07-high; checkpoint=gpt-5	Not reported
SWE-rebench · 2025-10 · toolsSWE-rebench · SWE-rebench standardized ReAct tools Benchmark details Average resolved rate across repeated runs, not best-of-N success. Disjoint complete calendar months within each model’s tested range; standardized harness and all-language task set. Unit: percent. Source collected 2026-10-18.	25.97% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-10-tools; gpt-5-2025-08-07-minimal; checkpoint=gpt-5	Not reported	42.35% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-10-tools; gpt-5-2025-08-07-medium; checkpoint=gpt-5	38.82% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-10-tools; gpt-5-2025-08-07-high; checkpoint=gpt-5	Not reported
SWE-rebench · 2025-11 · toolsSWE-rebench · SWE-rebench standardized ReAct tools Benchmark details Average resolved rate across repeated runs, not best-of-N success. Disjoint complete calendar months within each model’s tested range; standardized harness and all-language task set. Unit: percent. Source collected 2026-10-18.	34.89% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-11-tools; gpt-5-2025-08-07-minimal; checkpoint=gpt-5	Not reported	58.72% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-11-tools; gpt-5-2025-08-07-medium; checkpoint=gpt-5	55.74% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2025-11-tools; gpt-5-2025-08-07-high; checkpoint=gpt-5	Not reported

What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 8,916 results across 259 models · Collected 2026-10-18

GPT-5

Compare with

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context

Capability score evidence

Benchmark results