RankingEffort benchmarks

GPT-5.5 · XHigh

Compare models

169 measured results · 7 effort settings · 78 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-5.5 · XHigh5 of 7 capabilities supported

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

Agentic
A55.1Supported
Hard reasoning
A59.1Supported
Coding
A65.8Supported
Human pref
Unknown
Knowledge
A52.2Supported
Multimodal
A35.7Supported
Long context
A63.9Preliminary

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-5.5 · XHigh: 56.9 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?

Moonshot AI · 23.9% direct comparison weight · provider report
LiveBench · 8.1% direct comparison weight · independent of the model provider
Cognition · 8.1% direct comparison weight · independent of the model provider
Terminal-Bench · 8.1% direct comparison weight · independent of the model provider
Vals AI · 8.1% direct comparison weight · independent of the model provider
ARC Prize · 7.9% direct comparison weight · independent of the model provider
Artificial Analysis · 7.8% direct comparison weight · independent of the model provider
Scale · 7.6% direct comparison weight · independent of the model provider
SWE-rebench · 7.6% direct comparison weight · independent of the model provider
Epoch AI · 6.9% direct comparison weight · independent of the model provider
Datacurve · 5.9% direct comparison weight · independent of the model provider

Which benchmark families carry the weight?

Terminal-Bench · 8.6% direct comparison weight
DeepSWE · 8.4% direct comparison weight
FrontierCode · 8.1% direct comparison weight
ARC-AGI · 7.9% direct comparison weight
swe-rebench · 7.6% direct comparison weight
Other families (42) · 59.5% 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.

DeepSWE v1.1 review: Epoch AI reported grading errors in at least 23 of 113 tasks. These published results remain included. Review their limitations in Epoch's review when interpreting this evidence.

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	58 → 57	Up 1	203 remaining / 241 lost	57.8 (+0.9 points)
ARC Prize	60 → 62	Down 2	444 remaining / 0 lost	59.4 (+2.5 points)
Datacurve	60 → 58	Up 2	444 remaining / 0 lost	52.2 (-4.7 points)
LiveCodeBench	60 → 60	No movement	444 remaining / 0 lost	56.9 (-0.0 points)
Scale	60 → 60	No movement	444 remaining / 0 lost	58.9 (+2.0 points)
Terminal-Bench	60 → 61	Down 1	444 remaining / 0 lost	56.4 (-0.5 points)
Cognition	60 → 50	Up 10	444 remaining / 0 lost	57.0 (+0.1 points)
LiveBench	60 → 62	Down 2	443 remaining / 1 lost	57.2 (+0.4 points)
SWE-rebench	60 → 63	Down 3	444 remaining / 0 lost	54.9 (-2.0 points)
Epoch AI	60 → 60	No movement	431 remaining / 13 lost	57.1 (+0.3 points)
Vals AI	60 → 57	Up 3	443 remaining / 1 lost	58.1 (+1.3 points)
Provider reports · Anthropic	60 → 60	No movement	444 remaining / 0 lost	56.9 (-0.0 points)
Provider reports · Moonshot AI	60 → 44	Up 16	444 remaining / 0 lost	64.5 (+7.6 points)
Provider reports · Meta	60 → 60	No movement	444 remaining / 0 lost	56.8 (-0.0 points)
Provider reports · Mistral	60 → 60	No movement	444 remaining / 0 lost	56.9 (+0.0 points)
Provider reports · OpenAI	60 → 60	No movement	444 remaining / 0 lost	56.7 (-0.2 points)
All provider reports	60 → 47	Up 13	444 remaining / 0 lost	64.1 (+7.2 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
deepswe	51.7	-5.2
swe-rebench	54.9	-2.0
arc-agi	58.6	+1.7
hle	58.0	+1.2
gpqa	55.8	-1.0
swe-atlas-test-writing	57.7	+0.8
simpleqa	57.7	+0.8
swe-atlas-refactoring	56.2	-0.6
mmmu-pro	57.5	+0.6
frontiermath	56.3	-0.5
mmlu-pro	57.4	+0.5
livebench-instructions	57.4	+0.5
swe-atlas-qna	56.4	-0.5
livebench-coding	56.4	-0.5
vals-code-migration	56.5	-0.3
epoch-game-puzzles	56.6	-0.3
vals-legal-research	56.6	-0.3
swe-bench-pro	56.6	-0.2
livebench-math	57.1	+0.2
vals-excel-modeling	56.7	-0.2
arc-agi-3	56.7	-0.2
livebench-reasoning	56.7	-0.2
omniscience	57.0	+0.2
vals-finance-agent	57.0	+0.1
frontiercode	57.0	+0.1
tau-bench	56.8	-0.1
automationbench	57.0	+0.1
gdp-pdf	57.0	+0.1
analyst-agent	56.8	-0.1
ifbench	56.8	-0.1
itbench	56.8	-0.1
terminal-bench	56.8	-0.1
livebench-data	56.8	-0.1
enterprise-ops	56.9	+0.0
enigma-eval	56.9	+0.0
livebench-language	56.8	-0.0
aime	56.8	-0.0
harvey	56.9	+0.0
vals-tax-agent	56.9	+0.0
aa-lcr	56.9	+0.0
apex-agents	56.9	+0.0
gdpval	56.9	+0.0
terminal-bench-science	56.9	+0.0
critpt	56.9	+0.0
multi-swe-bench	56.9	-0.0
vibe-code	56.9	-0.0
gmmlu	56.9	+0.0
livecodebench	56.9	-0.0
briefcase	56.9	+0.0
math500	56.9	+0.0

What evidence is missing from the fit?

Across all collected settings for this model: 326 observations, 162 contributing, 0 matched without graph weight and 164 excluded. Counts do not establish rank eligibility.

9 · Benchmark family or source protocol has not been reviewed
3 · Original SciCode grading defects are under review; retained as supporting evidence, excluded from aggregate and capability scores. Corrected SciCode-Verified requires a separate protocol review. Sources: https://arxiv.org/abs/2608.04975 and https://artificialanalysis.ai/evaluations/scicode
1 · Fewer than two distinct model identities remain in this evaluation
7 · Evaluator has not been reviewed for independent admission
1 · Effort is not identified
7 · Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites GLM release blog or Vals AI, per model; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites GLM release blog, Artificial Analysis or OpenAI, per model; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites FrontierSWE leaderboard; retain as published context, not a new Moonshot comparison. Historical provider claim; secondary and superseded for current FrontierSWE comparisons by the independent Proximal FrontierSWE v2 mean@5 board at https://www.frontierswe.com/. Version-unspecified claims are not v2 results. Excluded from weighted Overall.
56 · Supporting evidence outside the reviewed capability core
1 · Kimi K3 Evaluation Details cites official PostTrainBench results; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites Anthropic or OpenAI release reports; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites official Agents Last Exam leaderboard; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites Artificial Analysis / APEX-Agents leaderboard; retain as published context, not a new Moonshot comparison.
3 · Kimi K3 Evaluation Details cites Vals AI; retain as published context, not a new Moonshot comparison.
47 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately
17 · Narrow specialist biological, medical or security evaluation; preserved as raw evidence outside general capability scoring.
1 · MiniMax cites this comparator from the official Terminal-Bench leaderboard; a common protocol with its internal Terminus 2 run is not established. See MiniMax M3 benchmark figure, Terminal-bench 2.1 methodology.
5 · Composite index or publisher-specific internal evaluation lacks a stable general-capability comparison unit.

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	Low	Medium	High	XHigh	None	Non-reasoning	Unspecified
AA-AnalystAgentArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of tasks solved on all 5 of 5 attempts · 80 tasks, 5 attempts per task · 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	50% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	7.5% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	1,000 Elo Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	1,088.3 Elo Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	1,136.33 Elo Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	Not reported
APEX-Agents-AAArtificial 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	Not reported	37.68% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	Not reported
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.	81% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	83% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	84.33% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	84.33% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	64% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	40.07% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	44.35% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	47.3% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	Not reported
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.	8% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	18.57% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	25.43% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	27.14% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	1.43% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
EnterpriseOps-Gym-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Overall task success rate across all eight enterprise domains (oracle tool mode) · Benchmark developed by ServiceNow Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	44.11% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	Not reported	46.64% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	Not reported
GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	20.6% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	22.6% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	21.2% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	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.	1,044.57 Elo Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	1,231 Elo Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	1,326.23 Elo Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	1,350.71 Elo Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	969.37 Elo Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	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.	91.01% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	92.63% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	93.23% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	93.54% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	76.77% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
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.	32.67% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	42.35% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	45.04% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	45.78% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	13.72% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
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.	64.35% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	70.95% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	71.63% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	75.85% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	46.12% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
ITBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Combined public and private Site Reliability Engineering (SRE) tasks · 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	45.81% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	Not reported
MLCR-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of tasks judged accurate, complete, and concise · Expert + compound (hard) sets · Benchmark developed by Wisedocs · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	11.11% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	Not reported	Not reported	17.78% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	15% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
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.	79.02% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	81.16% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	81.1% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	79.88% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	71.39% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	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.	15.1 Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	18.1 Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	18.75 Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	20.52 Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	-4.75 Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	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.	54.85% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	56.8% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	57.03% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	57.95% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	45.63% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark curated by scientists across 16 disciplines. SciCode is under review following an independent audit of the dataset · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	54.51% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	56.13% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	55.79% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	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.	83.92% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	91.81% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	92.98% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	93.86% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	69.3% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
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.	24.95% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	29.9% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	36.7% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	38.97% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	14.85% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	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.	52.27% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	57.58% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	59.85% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	60.61% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	49.24% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
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.	65.54% Reported settings GPT-5.5 (Low); Model source: https://artificialanalysis.ai/models/gpt-5-5-low	80.52% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	79.4% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	84.27% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	61.05% Reported settings GPT-5.5 (Non-reasoning); Model source: https://artificialanalysis.ai/models/gpt-5-5-non-reasoning	Not reported
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours Benchmark details Benchmark developed by the Laude Institute, Stanford, 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	5.05% Reported settings GPT-5.5 (Medium); Model source: https://artificialanalysis.ai/models/gpt-5-5-medium	9.09% Reported settings GPT-5.5 (High); Model source: https://artificialanalysis.ai/models/gpt-5-5-high	14.65% Reported settings GPT-5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/gpt-5-5	Not reported	Not reported	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.	33.33% Reported settings GPT-5.5 (Low); Source model ID: gpt-5-5-2026-04-22-thinking-low; 120 tasks	70.42% Reported settings GPT-5.5 (Medium); Source model ID: gpt-5-5-2026-04-22-thinking-medium; 120 tasks	83.33% Reported settings GPT-5.5 (High); Source model ID: gpt-5-5-2026-04-22-thinking-high; 120 tasks	85% Reported settings GPT-5.5 (XHigh); Source model ID: gpt-5-5-2026-04-22-thinking-xhigh; 120 tasks	Not reported	Not reported	Not reported
DeepSWE v1.1Datacurve · mini-swe-agent Benchmark details Original public board. Configuration and task coverage may differ; inspect source. Unit: percent. Source collected 2026-10-18.	27% Reported settings gpt-5-5; Reported effort: low; 113 unique tasks; 452 attempts; 4 runs	54% Reported settings gpt-5-5; Reported effort: medium; 113 unique tasks; 452 attempts; 4 runs	64.4% Reported settings gpt-5-5; Reported effort: high; 113 unique tasks; 452 attempts; 4 runs	67% Reported settings gpt-5-5; Reported effort: xhigh; 113 unique tasks; 452 attempts; 4 runs	Not reported	Not reported	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-5.5 · XHigh

Compare with

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context

Capability score evidence

Benchmark results