RankingGPT-5.5

GPT-5.5

Compare models

Data updated 22 Nov 2026

GPT-5.5 by OpenAI is a gpt-5.5 model. This profile collects the published benchmark results and reported effort settings for this model. Configurations are compared on the effort page. Ranks are on the leaderboard.

112 published benchmark measures · 158 measured effort results · 23 benchmark families contribute across 6 task areas. 5 capability estimates available in the full profile. See all results ↓ · Compare published benchmarks →

Model evidence summary

This profile combines published settings. It is not a runnable configuration or a leaderboard rank. Compare measured configurations →

Performance profile

Capabilities

Adjusted comparison score · 0–100
GPT-5.5 capability radarComparison scores from 0 to 100. Unknown capabilities have no point; lines stop at gaps. Hollow points are preliminary. Exact values and support labels follow the chart.AgenticHard reasoningCodingHuman prefKnowledgeMultimodalLong context50100GPT-5.5 · Agentic: 31.5 · SupportedGPT-5.5 · Hard reasoning: 55.2 · SupportedGPT-5.5 · Coding: 38.2 · SupportedGPT-5.5 · Multimodal: 45.5 · SupportedGPT-5.5 · Long context: 38.5 · Supported

Filled points are supported; hollow points are preliminary. Lines stop at unknown capabilities. Exact values and sources follow below.

Agentic11 families · 0 with independent evidence · Supported31.5

11 core families; 16 direct opponents across 7 labs.

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: 27.7–44.1; 0/13 scenarios unsupported. Without one publisher: 23.9–45.5; 0/15 unsupported. Smoothing check: 28.5–37.0; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage; source sensitivity still requires inspection. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

Hard reasoning4 families · 1 with independent evidence · Supported55.2

4 core families; 60 direct opponents across 10 labs.

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: 40.4–66.6; 0/6 scenarios unsupported. Without one publisher: 36.4–58.9; 0/16 unsupported. Smoothing check: 53.6–56.3; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

Coding3 families · 1 with independent evidence · Supported38.2

3 core families; 32 direct opponents across 10 labs.

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: 29.2–53.2; 0/6 scenarios unsupported. Without one publisher: 35.4–41.5; 0/19 unsupported. Smoothing check: 37.5–40.6; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

Human pref1 families · 1 with independent evidence · PreliminaryUnknown

1 core families; 75 direct opponents across 17 labs. No connected comparison to the complete reference panel

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 1/1 scenarios unsupported. Without one publisher: No supported estimate; 1/1 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: cross-family and cross-publisher predictions could not be evaluated from the available comparison network. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

  • Human preference (Arena): 89.3 observed win share
    lmarena.ai · Source 1
KnowledgeNo comparable evidenceUnknown

0 core families; 0 direct opponents across 0 labs. No admitted core comparisons

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 4/4 scenarios unsupported. Without one publisher: No supported estimate; 5/5 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: almost no publisher-held-out comparisons remained estimable. Cross-source validity is unresolved. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

    Multimodal2 families · 0 with independent evidence · Supported45.5

    2 core families; 12 direct opponents across 6 labs.

    Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

    Refit without one family: 45.5–45.6; 2/8 scenarios unsupported. Without one publisher: 41.1–55.3; 0/8 unsupported. Smoothing check: 45.5–46.2; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

    Audit: limited publisher-held-out coverage. Cross-source validity is unresolved. Validation audit.

    Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

    Long context2 families · 0 with independent evidence · Supported38.5

    2 core families; 6 direct opponents across 4 labs.

    Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

    Refit without one family: 38.5–39.3; 2/5 scenarios unsupported. Without one publisher: 38.5–40.3; 2/5 unsupported. Smoothing check: 37.0–41.3; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

    Audit: family and publisher holdouts underperformed neutral predictions. Treat this as an exploratory estimate. Validation audit.

    Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

    Scores estimate outcomes against a shared reference panel; they are not accuracy percentages. Sparse or disconnected evidence cannot qualify an overall profile. Open a capability to inspect its evidence.

    Reported effort · XHigh + unspecified

    Settings reported in this model's published benchmark results, including results outside the aggregate. Effort names are provider-specific. These are not API defaults or equal compute budgets.

    Mixed settings means multiple settings occur in the evidence. Best across efforts means the source selected its best reported result across settings; it does not mean Max. Unspecified settings stay unknown. This model-summary chart combines reported settings. The leaderboard keeps identified configurations separate and excludes unknown effort.

    Inspect each result and its source ↓ · Download effort evidence
    Score contributions and missing evidence

    23 contributing families across 6 capabilities. Fixed reference panels do not change when the catalog expands.

    Capability is fitted jointly across families. Capability estimates below describe different task areas; their weighted sum is not the Capability score.

    Results without reviewed compatibility or a reference match remain in the raw evidence below. Coverage counts only contributing results.

    Model information & shareable badge
    Lab
    OpenAI
    Catalog status
    active
    Availability
    Public provider catalog; account and region restrictions may apply
    Family
    gpt-5.5
    Released
    —
    Context
    —
    License
    proprietary
    Model card
    https://developers.openai.com/api/docs/models/all
    Default Capability family coverage
    Documented-evidence family coverage/badge/gpt-5.5.svg

    GPT-5.5 measured results by effort setting

    Each row is one published result for one effort setting. Evidence collected 2026-11-22. Compare the effort settings →

    BenchmarkEffortScoreSource
    AA-AnalystAgentNon-reasoning7.5%Artificial Analysis
    AA-AnalystAgentXHigh50.0%Artificial Analysis
    AA-Briefcase EloHigh1087Artificial Analysis
    AA-Briefcase EloMedium1000Artificial Analysis
    AA-Briefcase EloXHigh1135Artificial Analysis
    AA-LCR v1.1High84.3%Artificial Analysis
    AA-LCR v1.1Low81.0%Artificial Analysis
    AA-LCR v1.1Medium83.0%Artificial Analysis
    AA-LCR v1.1Non-reasoning64.0%Artificial Analysis
    AA-LCR v1.1XHigh84.3%Artificial Analysis
    Agents' Last ExamHigh23.0%Agents' Last Exam
    Agents' Last ExamHigh21.7%Agents' Last Exam
    Agents' Last ExamHigh19.7%Agents' Last Exam
    Agents' Last ExamMedium21.4%Agents' Last Exam
    Agents' Last ExamXHigh26.6%Agents' Last Exam
    APEX-Agents-AAXHigh37.7%Artificial Analysis
    ARC-AGI-2High83.3%ARC Prize
    ARC-AGI-2Low33.3%ARC Prize
    ARC-AGI-2Medium70.4%ARC Prize
    ARC-AGI-2XHigh85.0%ARC Prize
    AutomationBench-AAHigh44.3%Artificial Analysis
    AutomationBench-AAMedium40.1%Artificial Analysis
    AutomationBench-AAXHigh47.3%Artificial Analysis
    CritPtHigh25.4%Artificial Analysis
    CritPtLow8.0%Artificial Analysis
    CritPtMedium18.6%Artificial Analysis
    CritPtNon-reasoning1.4%Artificial Analysis
    CritPtXHigh27.1%Artificial Analysis
    DeepSWE v1.1High64.4%Datacurve
    DeepSWE v1.1Low27.0%Datacurve
    DeepSWE v1.1Medium54.0%Datacurve
    DeepSWE v1.1XHigh67.0%Datacurve
    EnterpriseOps-Gym-AAMedium44.1%Artificial Analysis
    EnterpriseOps-Gym-AAXHigh46.6%Artificial Analysis
    Epoch · Chess Puzzles · v2.0.0Low30.0%Epoch AI
    Epoch · Chess Puzzles · v2.0.0None9.0%Epoch AI
    Epoch · Chess Puzzles · v2.0.0XHigh63.0%Epoch AI
    Epoch · FrontierMath Tier 4 · v2.0.0XHigh72.5%Epoch AI
    Epoch · FrontierMath Tiers 1–3 · v2.0.0XHigh85.3%Epoch AI
    Epoch · GPQA diamond · v1.0.11None77.3%Epoch AI
    Epoch · GPQA diamond · v1.0.8Low90.7%Epoch AI
    Epoch · Mystery Game Puzzles · v1.0.4 · Tool submissionHigh52.0%Epoch AI
    Epoch · Mystery Game Puzzles · v1.0.4 · Tool submissionLow28.0%Epoch AI
    Epoch · Mystery Game Puzzles · v1.0.4 · Tool submissionNone18.0%Epoch AI
    Epoch · Mystery Game Puzzles · v1.0.4 · Tool submissionXHigh56.0%Epoch AI
    Epoch · OTIS Mock AIME 2024-2025 · v1.1.6Low84.4%Epoch AI
    Epoch · OTIS Mock AIME 2024-2025 · v1.1.6None57.8%Epoch AI
    Epoch · SimpleQA Verified · v1.2.0XHigh63.0%Epoch AI
    FrontierCode 1.1 · extended-codexHigh54.0%Cognition
    FrontierCode 1.1 · extended-codexLow45.6%Cognition
    FrontierCode 1.1 · extended-codexMedium51.4%Cognition
    FrontierCode 1.1 · extended-codexXHigh56.7%Cognition
    FrontierCode 1.1 · main-codexHigh40.6%Cognition
    FrontierCode 1.1 · main-codexLow30.6%Cognition
    FrontierCode 1.1 · main-codexMedium36.6%Cognition
    FrontierCode 1.1 · main-codexXHigh43.0%Cognition
    FrontierMath ErdősXHigh0.0%Epoch AI
    GDP.pdf: All-pass RateHigh22.6%Artificial Analysis
    GDP.pdf: All-pass RateMedium20.6%Artificial Analysis
    GDP.pdf: All-pass RateXHigh21.2%Artificial Analysis
    GDPval-AA v2.1 LeaderboardHigh1335Artificial Analysis
    GDPval-AA v2.1 LeaderboardLow1053Artificial Analysis
    GDPval-AA v2.1 LeaderboardMedium1242Artificial Analysis
    GDPval-AA v2.1 LeaderboardNon-reasoning978Artificial Analysis
    GDPval-AA v2.1 LeaderboardXHigh1357Artificial Analysis
    GPQA DiamondHigh93.2%Artificial Analysis
    GPQA DiamondLow91.0%Artificial Analysis
    GPQA DiamondMedium92.6%Artificial Analysis
    GPQA DiamondNon-reasoning76.8%Artificial Analysis
    GPQA DiamondXHigh93.5%Artificial Analysis
    Humanity's Last ExamHigh45.0%Artificial Analysis
    Humanity's Last ExamLow32.7%Artificial Analysis
    Humanity's Last ExamMedium42.4%Artificial Analysis
    Humanity's Last ExamNon-reasoning13.7%Artificial Analysis
    Humanity's Last ExamXHigh45.8%Artificial Analysis
    IFBenchHigh71.6%Artificial Analysis
    IFBenchLow64.4%Artificial Analysis
    IFBenchMedium71.0%Artificial Analysis
    IFBenchNon-reasoning46.1%Artificial Analysis
    IFBenchXHigh75.9%Artificial Analysis
    ITBench-AAXHigh45.8%Artificial Analysis
    LiveBench · AMPS Hard · 2026-06-25XHigh97.0%LiveBench
    LiveBench · code completion · 2026-06-25XHigh82.6%LiveBench
    LiveBench · code generation · 2026-06-25XHigh81.7%LiveBench
    LiveBench · connections · 2026-06-25XHigh100.0%LiveBench
    LiveBench · consecutive events · 2026-06-25XHigh88.8%LiveBench
    LiveBench · integrals with game · 2026-06-25XHigh99.0%LiveBench
    LiveBench · javascript · 2026-06-25XHigh63.6%LiveBench
    LiveBench · logic with navigation · 2026-06-25XHigh76.0%LiveBench
    LiveBench · math comp · 2026-06-25XHigh96.1%LiveBench
    LiveBench · olympiad · 2026-06-25XHigh91.4%LiveBench
    LiveBench · paraphrase · 2026-06-25XHigh71.7%LiveBench
    LiveBench · plot unscrambling · 2026-06-25XHigh74.1%LiveBench
    LiveBench · python · 2026-06-25XHigh55.0%LiveBench
    LiveBench · simplify · 2026-06-25XHigh66.3%LiveBench
    LiveBench · spatial · 2026-06-25XHigh98.0%LiveBench
    LiveBench · story generation · 2026-06-25XHigh72.5%LiveBench
    LiveBench · summarize · 2026-06-25XHigh72.3%LiveBench
    LiveBench · tablejoin · 2026-06-25XHigh55.9%LiveBench
    LiveBench · tablereformat · 2026-06-25XHigh100.0%LiveBench
    LiveBench · theory of mind · 2026-06-25XHigh84.6%LiveBench
    LiveBench · typescript · 2026-06-25XHigh43.3%LiveBench
    LiveBench · typos · 2026-06-25XHigh88.0%LiveBench
    LiveBench · zebra puzzle · 2026-06-25XHigh100.0%LiveBench
    LMArena Agent ArenaUnspecified1.0LMArena
    LMArena Agent ArenaXHigh4.2LMArena
    MLCR-AAHigh25.0%Artificial Analysis
    MLCR-AAMedium22.2%Artificial Analysis
    MLCR-AAXHigh28.9%Artificial Analysis
    MMMU-ProHigh81.1%Artificial Analysis
    MMMU-ProLow79.0%Artificial Analysis
    MMMU-ProMedium81.2%Artificial Analysis
    MMMU-ProNon-reasoning71.4%Artificial Analysis
    MMMU-ProXHigh79.9%Artificial Analysis
    SciCodeHigh56.1%Artificial Analysis
    SciCodeMedium54.5%Artificial Analysis
    SciCodeXHigh55.8%Artificial Analysis
    SWE Atlas · QnA · codexXHigh45.4%Scale
    SWE Atlas · Refactoring · codexXHigh44.8%Scale
    SWE Atlas · Test Writing · codexXHigh42.6%Scale
    SWE-rebench · 2026-03 · toolsMedium60.4%SWE-rebench
    SWE-rebench · 2026-03 · toolsXHigh63.0%SWE-rebench
    SWE-rebench · 2026-04 · toolsMedium57.2%SWE-rebench
    SWE-rebench · 2026-04 · toolsXHigh60.5%SWE-rebench
    Terminal-Bench 2.1XHigh83.2%Terminal-Bench
    Terminal-Bench 2.1XHigh78.0%Terminal-Bench
    Terminal-Bench HardHigh59.8%Artificial Analysis
    Terminal-Bench HardLow52.3%Artificial Analysis
    Terminal-Bench HardMedium57.6%Artificial Analysis
    Terminal-Bench HardNon-reasoning49.2%Artificial Analysis
    Terminal-Bench HardXHigh60.6%Artificial Analysis
    Terminal-Bench v2.1High79.4%Artificial Analysis
    Terminal-Bench v2.1Low65.5%Artificial Analysis
    Terminal-Bench v2.1Medium80.5%Artificial Analysis
    Terminal-Bench v2.1Non-reasoning61.0%Artificial Analysis
    Terminal-Bench v2.1XHigh84.3%Artificial Analysis
    Terminal-Bench v4.0High9.1%Artificial Analysis
    Terminal-Bench v4.0Medium5.1%Artificial Analysis
    Terminal-Bench v4.0XHigh14.6%Artificial Analysis
    Toolathlon-VerifiedXHigh73.5%HKUST NLP
    Vals AI · Code Migration v1XHigh45.2%Vals AI
    Vals AI · Excel Modeling Benchmark v1XHigh64.5%Vals AI
    Vals AI · Finance Agent v2XHigh51.8%Vals AI
    Vals AI · Legal Research Bench v1XHigh40.4%Vals AI
    Vals AI · MMLU-Pro v1XHigh88.1%Vals AI
    Vals AI · Tax Agent Bench v1: Weighted PassXHigh60.5%Vals AI
    Vals AI · Terminal-Bench 2.1High76.4%Vals AI
    Vals AI · Vibe Code Bench v1.1XHigh69.8%Vals AI
    𝜏²-Bench TelecomHigh93.0%Artificial Analysis
    𝜏²-Bench TelecomLow83.9%Artificial Analysis
    𝜏²-Bench TelecomMedium91.8%Artificial Analysis
    𝜏²-Bench TelecomNon-reasoning69.3%Artificial Analysis
    𝜏²-Bench TelecomXHigh93.9%Artificial Analysis
    𝜏³-BankingHigh36.7%Artificial Analysis
    𝜏³-BankingLow24.9%Artificial Analysis
    𝜏³-BankingMedium29.9%Artificial Analysis
    𝜏³-BankingNon-reasoning14.8%Artificial Analysis
    𝜏³-BankingXHigh39.0%Artificial Analysis
    Benchmark scores & sources

    Original results, evaluation harnesses, and evidence behind this model.

    BenchmarkBucketScoreHarnessEvidenceSource-recorded date
    WorldVQA ForceAnswer · source release snapshot; version not specifiedSupporting evidence38.5%
    Reported settings & source

    Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

    First-party reported result.

    moonshotai/Kimi-K3 · Performance table: WorldVQA ForceAnswer · reviewed 2026-11-20

    Published configuration

    Effort: XHigh

    lab self-report

    Supporting evidence outside the reviewed capability core

    Reviewed 2026-11-20
    YC-Bench · source release snapshot; version not specifiedSupporting evidence1300000 USD
    Reported settings & source

    Launch figure; agent final assets

    First-party reported result; comparator results retain the source evaluation setup.

    MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

    Published configuration

    Effort: Not specified

    lab self-report

    Supporting evidence outside the reviewed capability core

    Reviewed 2026-09-06
    ZeroBench (pass@5) · source release snapshot; version not specifiedSupporting evidence22%
    Reported settings & source

    Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. without tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

    First-party reported result.

    moonshotai/Kimi-K3 · Performance table: ZeroBench (pass@5) · reviewed 2026-11-20

    Published configuration

    Effort: XHigh

    lab self-report

    Supporting evidence outside the reviewed capability core

    Reviewed 2026-11-20
    ZeroBench (pass@5) · source release snapshot; version not specifiedSupporting evidence41%
    Reported settings & source

    Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. with tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

    First-party reported result.

    moonshotai/Kimi-K3 · Performance table: ZeroBench (pass@5) · reviewed 2026-11-20

    Published configuration

    Effort: XHigh

    lab self-report

    Supporting evidence outside the reviewed capability core

    Reviewed 2026-11-20
    τ³-Banking · source release snapshot; version not specifiedSupporting evidence31.3%
    Reported settings & source

    Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

    Cited result; see benchmark-specific Evaluation Details.

    Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

    moonshotai/Kimi-K3 · Performance table: τ³-Banking · reviewed 2026-11-20

    Published configuration

    Effort: XHigh

    lab self-report

    Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

    Reviewed 2026-11-20
    GDPval-AAAgentic————
    GPQA DiamondHard reasoning————
    Humanity's Last ExamHard reasoning————
    LiveCodeBenchCoding————
    MMLU-ProKnowledge————
    OSWorld-VerifiedAgentic————
    SWE-bench ProAgentic————
    SWE-bench VerifiedAgentic————