OpenAI, released Mar 5, 2026

GPT-5.4price, context, benchmarks and release details

100% confidence 100 percent, Full confidence, 7 of 8 expected sources in
69.7
SI Score
#13 of 142 ranked models
Input, per 1M tokens
$2.50models.devFirst-party hosted API; MIT models.dev transcription. Provider documentation: https://platform.openai.com/docs/models. Exact canonical endpoint; lowest short-context Standard USD token tier; cache/batch discounts excluded. Deprecated endpoints excluded; openai/gpt-5.4 Retrieved Oct 9, 2026 · MIT
Open source ↗
Output, per 1M tokens
$15.00models.devFirst-party hosted API; MIT models.dev transcription. Provider documentation: https://platform.openai.com/docs/models. Exact canonical endpoint; lowest short-context Standard USD token tier; cache/batch discounts excluded. Deprecated endpoints excluded; openai/gpt-5.4 Retrieved Oct 9, 2026 · MIT
Open source ↗
Context window
1.1Mmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
Max output
128Kmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
Released
Mar 5, 2026OpenAI API changelogPublished source fact Retrieved Oct 9, 2026 · factual citation
Open source ↗

How the score breaks down

Coding (weight 40 percent) 64.9
Math (weight 15 percent) 78.5
Preference (weight 15 percent) 77.9
Reasoning (weight 30 percent) 71.6

Weights: reasoning 30%, math 15%, coding 40%, preference 15%. Results use fixed 0–100 scales before averaging, and thin evidence is pulled toward 50. Method si-v3-retained-evidence-2, computed Oct 9, 2026, 05:13 UTC.

Around it on the leaderboard

  1. 11 Muse Spark 1.3 70.1
  2. 12 Gemini 3.7 Flash 70.0
  3. 13 GPT-5.4 69.7
  4. 14 GPT-5.5 69.5
  5. 15 Claude Sonnet 5.5 69.4

Full leaderboard

Benchmark results

35 benchmarks, 57 results

Each row shows the best published result. Where a model was tested at several settings, such as reasoning effort, open the row to see each one. Hover or tap a value for its source.

ARC-AGI-1 (public eval)reasoning 96.4%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 96.4 4 settings
  • Setting 1 95.6%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 2 80%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 3 92%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 4 96.4%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-1 (public eval)
ARC-AGI-1 (semi-private)reasoning 93.7%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 93.7 4 settings
  • Setting 1 92.7%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 2 68.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 3 86.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 4 93.7%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-1 (semi-private)
ARC-AGI-2reasoning 73.3%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] VerifiedPublished Apr 23, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 73.3
ARC-AGI-2 (public eval)reasoning 84.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 84.2 4 settings
  • Setting 1 75.8%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 2 23.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 3 58.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 4 84.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-2 (public eval)
ARC-AGI-2 (semi-private)reasoning 74.0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 74.0 4 settings
  • Setting 1 67.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 2 29.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 3 55.4%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • Setting 4 74.0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] gpt-5-4-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-2 (semi-private)
ARC-AGI-3 (semi-private)reasoning 0.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] openai-gpt-5-4-2026-03-05-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 0.2
GPQA Diamondreasoning 93.3%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] xhighPublished Mar 6, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 93.3 6 settings
  • no reasoning 74.7%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] nonePublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • low effort 84.8%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] lowPublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • medium effort 88.9%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] mediumPublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • high effort 89.9%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] highPublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • xhigh effort 93.3%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] xhighPublished Mar 6, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • Setting 6 92.8%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] Published Apr 23, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
    Open source ↗
About GPQA Diamond
Humanity's Last Examreasoning 39.8%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] no toolsPublished Apr 23, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 39.8
Humanity's Last Exam (Scale AI)reasoning 36.2%Humanity’s Last ExamPotential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. [variant] Published Mar 10, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 36.2
Humanity's Last Exam (with tools)reasoning 52.1%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] with toolsPublished Apr 23, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 52.1
LiveBench Reasoning: connectionsreasoning 100%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 100.0
LiveBench Reasoning: consecutive eventsreasoning 86.2%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 86.2
LiveBench Reasoning: logic with navigationreasoning 68%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 68.0
LiveBench Reasoning: spatialreasoning 98%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 98.0
LiveBench Reasoning: theory of mindreasoning 88.5%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 88.5
LiveBench Reasoning: zebra puzzlesreasoning 98%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 98.0
FrontierMath Tier 4math 27.1%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] Tier 4Published Apr 23, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 27.1
FrontierMath Tier 4 (v2)math 49%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] xhighPublished Jun 11, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 49.0
FrontierMath Tiers 1–3math 47.6%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] Tier 1-3Published Apr 23, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 47.6
FrontierMath Tiers 1–3 (v2)math 78.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] xhighPublished Jun 11, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 78.6
LiveBench Math: AMPS Hardmath 98%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 98.0
LiveBench Math: competition mathmath 94.1%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 94.1
LiveBench Math: integralsmath 93%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 93.0
LiveBench Math: olympiadmath 91.5%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 91.5
LiveBench Math: simplifymath 70%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 70.0
OTIS Mock AIME 2024–2025math 97.8%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] highPublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 97.8 5 settings
  • no reasoning 57.8%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] nonePublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • low effort 84.4%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] lowPublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • medium effort 95.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] mediumPublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • high effort 97.8%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] highPublished Jul 15, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • xhigh effort 95.3%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] xhighPublished Mar 6, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
About OTIS Mock AIME 2024–2025
LiveBench Coding: code completioncoding 80.4%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 80.4
LiveBench Coding: code generationcoding 74.6%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 74.6
LiveBench Coding: JavaScriptcoding 68.2%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 68.2
LiveBench Coding: Pythoncoding 50%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 50.0
LiveBench Coding: TypeScriptcoding 43.3%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 43.3
SWE-bench Pro (public)coding 59.1%Official model cards via models.devLab-reported; metric resolve rate; transcribed by MIT models.dev catalog; not independently evaluated [variant] public Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 59.1 2 settings
  • Setting 1 59.1%Official model cards via models.devLab-reported; metric resolve rate; transcribed by MIT models.dev catalog; not independently evaluated [variant] public Retrieved Oct 9, 2026 · factual citation; MIT transcription
    Open source ↗
  • Setting 2 43.4%SWE-bench Pro (public)Published steward score [variant] Published Apr 8, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About SWE-bench Pro (public)
SWE-bench Verifiedcoding 76.9%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] highPublished Mar 6, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 76.9
Terminal-Bench 2.0coding 75.1%Official model cards via models.devLab-reported; metric success rate; transcribed by MIT models.dev catalog; not independently evaluated [variant] 2.0Published Apr 23, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 75.1
LMArena Textpreference 1452.1 eloLMArena / ArenaPublished source fact [variant] text / overallPublished Oct 8, 2026 Retrieved Oct 9, 2026 · CC-BY-4.0
Open source ↗ 77.9

Normalization uses a fixed 0–100 scale for each unit, independent of other models. Compare evaluation conditions before reading a small gap as decisive. “Lab-reported” marks the provider's own published figure.

Details and sources

Open weights
Nomodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
License
not yet reported
Input modalities
text, image, pdfmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
First seen by SuperIndex
Oct 8, 2026
Coverage
94% of expected source weight

Reported (7)

  • ARC PrizeOct 8, 2026
  • Epoch AI BenchmarkingOct 8, 2026
  • Humanity’s Last ExamOct 8, 2026
  • Official model cards via models.devOct 8, 2026
  • LiveBenchOct 8, 2026
  • LMArena / ArenaOct 8, 2026
  • SWE-bench Pro (public)Oct 8, 2026

Awaiting (1)

  • Terminal-Bench6% of weight

Confidence rises as pending sources publish. Some sources never cover some models, so confidence reaches 100% at 80% of expected weight.

What changed

All releases · RSS feed

Alerts on this device

What to be alerted about
RSS feed