Inference
EN
OpenAI

GPT-5.5

GPT-5.5 is a reasoning model from OpenAI. 68 benchmarks count toward its index, in 7 categories.

  • Closed weights
  • Reasoning
  • Input: file · image · text
  • Released 2026-04
RankedEnough results for an index.
IndexAll categories, out of 100.69.1±2.5Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.52/sMedian 67/s
First tokenSeconds to the first output token.2.7 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$5.00batch $2.50Median $0.60
Output / 1MUS dollars per 1M output tokens.$30.00batch $15.00 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.1.05MMedian 262K
EloLMArena rating and rank.1466±4 #36Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
openai/gpt-5.5
Max outputMaximum output tokens per request.
128K
Knowledge cutoffLast date of training data.
1 Dec 2025
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.50
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
8 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
DatabricksCloud$5.00$30.001.05M
DigitalOceanCloud$5.00$30.001M
GMI CloudHost$5.00$30.001.05M
Microsoft AzureCloud$5.00$30.001.05M
OpenRouterRouter$5.00$30.001.05M
SAP AI CoreCloud$5.00$30.001.05M
Snowflake CortexCloud$5.00$30.001.05M
Amazon BedrockCloud$5.50$33.001M

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.25 counted65.842 / 218
CodingCode writing and repair.15 counted65.539 / 209
ReasoningLogic problems and puzzles.10 counted74.117 / 227
MultimodalTasks with images and text.4 counted66.820 / 84
KnowledgeFacts and expert knowledge.6 counted70.419 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.2 counted64.86 / 63
MathMath problems.6 counted74.716 / 168

Results

68 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
OTIS Mock AIME 2024-2025Math100.0%1 / 200
FrontierMath-2025-02-28-PrivateMath51.7% ±2.91 / 69
Terminal-Bench 2.0 · Vals AIAgentic73.2% ±4.01 / 66
Terminal-Bench 2.0 · BenchLMAgentic82.0%1 / 41
Artificial Analysis Long Context ReasoningReasoning84.3%7 / 217
Gert Labs Composite Game BenchmarkAgentic72.9%2 / 52
Artificial Analysis GPQA DiamondKnowledge93.5%8 / 199
τ²-Bench Tool-Agent-User EvaluationAgentic98.0%8 / 149
Graduate-Level Google-Proof Q&AKnowledge93.6%8 / 141
GPQA diamond · Epoch AIKnowledge94.0% ±1.513 / 217
Artificial Analysis Omniscience AccuracyKnowledge58.0%10 / 164
LiveBench Data AnalysisReasoning81.6%5 / 80
SWE-Bench verified · Epoch AICoding80.6% ±1.82 / 32
FrontierMath-Tier-4-2025-07-01-PrivateMath35.4% ±6.94 / 55
GPQA Diamond · Vals AIKnowledge93.2% ±1.310 / 135
Chess PuzzlesReasoning63.0% ±4.97 / 90
Critical Physics TasksReasoningAbove fitFar above this model's index.27.1%17 / 214
MMMU ProMultimodal88.3% ±0.88 / 92
GPQA Diamond · BenchLMKnowledge93.6%8 / 84
ARC-AGI-2 (semi-private)Reasoning85.0%13 / 131
Artificial Analysis IFBenchInstruction75.9%16 / 146
Artificial Analysis Humanity's Last ExamKnowledge45.8%26 / 222
Agent Arena command recoveryAgentic9.56 / 50
LiveBench LanguageKnowledge87.4%8 / 66
Terminal-Bench HardAgentic60.6%3 / 23
LiveBench CodingCoding82.1%9 / 66
SkillsBenchCoding62.2% ±4.55 / 35
Mystery Game PuzzlesReasoning56.0% ±5.012 / 80
ARC-AGI-1 (semi-private)Reasoning95.0%19 / 126
LiveBench MathematicsMath95.9%10 / 66
MMLU ProKnowledge88.1% ±0.321 / 135
SimpleQA VerifiedKnowledge63.0% ±1.516 / 91
FrontierMath-Tiers-1-3-v2-PrivateMath85.3% ±2.116 / 90
LiveBench ReasoningReasoning89.7%15 / 80
Massive Multi-discipline Multimodal Understanding ProMultimodal81.2%10 / 53
APEX-Agents-AAAgentic37.7%6 / 30
Artificial Analysis MMMU-ProMultimodal79.9%23 / 113
Terminal-Bench 2.1 · Vals AIAgentic76.4% ±0.616 / 75
τ²-bench BankingAgentic44.6%6 / 28
SWE-benchCoding82.6% ±1.719 / 87
FrontierMath-Tier-4-v2-PrivateMath72.5% ±7.116 / 72
ToolathlonAgentic55.6%5 / 22
ExtractBench: schema-guided enterprise document extraction (LlamaIndex leaderboard)Multimodal93.3%6 / 26
LiveCodeBenchCoding85.3% ±1.033 / 142
Code MigrationCoding45.2% ±4.218 / 77
Artificial Analysis SciCodeCoding55.8%27 / 114
Humanity's Last ExamKnowledge52.2%19 / 77
React Native EvalsCoding84.7%4 / 16
BrowseCompAgentic84.4%19 / 63
GDPval-AA normalizedAgentic42.5%44 / 133
LiveBench Long AgenticAgentic44.6%17 / 51
OSWorld-VerifiedAgentic78.7%13 / 37
Artificial Analysis AnalystAgentAgentic50.0%10 / 26
ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable jobAgentic20.0%7 / 18
LiveBench Instruction FollowingInstruction70.7%26 / 66
Vibe Code Bench v1.1Coding69.8% ±4.550 / 123
Furniture AssemblyReasoning44.2% ±6.416 / 37
LiveBench Coding 2.0Coding85.1%23 / 52
EBR-benchReasoning34.3% ±2.014 / 31
SWE-bench ProCoding58.6%38 / 83
MCP AtlasAgentic75.3%24 / 52
MazeBench: visual spatial reasoning in a 3D open worldAgentic0.0%23 / 48
Humanity's Last Exam without toolsKnowledge41.4%26 / 53
LiveBench Agentic CodingAgentic54.0%33 / 66
JobBenchAgentic42.7%18 / 36
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal66.4%20 / 39
Artificial Analysis ITBench-AAAgentic45.8%13 / 24
CyberGymAgentic81.8%16 / 29
Artificial Analysis Terminal-Bench v2.1Agentic84.3%12 / 21
LiveBench Browser UseAgentic48.0%30 / 49
FrontierCode 1.1 MainCoding43.0%11 / 17
FrontierCode 2.0Coding22.6%18 / 27
ARC-AGI-3 (semi-private)Reasoning0.4%18 / 26
cursorBench32Coding58.4%13 / 18
LiveBench DesignAgentic70.3%39 / 52
ExploitGymAgentic13.4%18 / 23
PostTrainBench v1.1Coding27.2%12 / 15
LiveBench ResearchAgentic61.6%42 / 52
OfficeQA ProMultimodal54.1%13 / 16
Artificial Analysis Tau3-BankingAgentic39.0%18 / 22
OSWorld 2.0Agentic13.0%21 / 25
MirrorCodeCoding23.3% ±10.133 / 39
CorpFin v2Finance68.4% ±0.99 / 131
LegalBenchLegal86.5% ±0.413 / 149
Finance Agent v1.1Finance60.0% ±2.85 / 51
MortgageTaxFinance68.8% ±0.910 / 97
ARC-AGI-2 (public eval)Reasoning90.6%14 / 122
CaseLaw v2Legal66.2% ±0.37 / 53
ARC-AGI-1 (public eval)Reasoning97.5%19 / 116
MedScribeHealthcare86.9% ±1.919 / 109
TaxEval v2Finance75.0% ±0.925 / 143
MedCodeHealthcare49.1% ±2.226 / 106
ProgramBenchCoding0.5% ±0.519 / 65
EMBFinance64.5% ±2.926 / 75
Legal Research BenchLegal40.4% ±3.435 / 90
Finance Agent (v2)Finance51.8% ±0.639 / 79
SRE BenchOther3.8% ±1.219 / 36
FrontierMath-ErdosMath0.0%8 / 15
Tax Agent BenchFinance21.8% ±3.049 / 82
Public Benefits Bench v1Public benefits57.2% ±1.38 / 13
Public Benefits Bench v1.1Public benefits60.9% ±1.332 / 50
OEIS OpenMath26.2% ±2.02 / 3
Time Horizon Index: KSPOther8.3%10 / 13
OEIS Open LiteMath36.0% ±4.84 / 5
OpenAI MRCR v2 8-needle 128K-256KReasoning87.5%—
MMMU-Pro with PythonMultimodal83.2%—
OpenAI MRCR v2 8-needle 64K-128KReasoning83.1%—
LifeVal: everyday life tasksAgentic77.4%—
cursorBench31Coding59.2%—
Artificial Analysis EnterpriseOps-GymAgentic46.6%—
BioBasicsBench: everyday wet-lab scienceReasoning35.4%—
ResearchClawBenchAgentic17.0%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYLiveBenchNo licence statedARC Prize FoundationTerms restrict reuseVals AINo licence statedSierra ResearchMITLMArenaCC BY 4.0models.devMIT

History

5 releases
Index70.6 → 69.1
23 Oct 202612 Nov 2026
Output / 1M$30.00 → $30.00
23 Oct 202612 Nov 2026

Questions

How much does GPT-5.5 cost?
$5.00 per 1M input tokens, $30.00 per 1M output tokens.
How good is GPT-5.5?
Its index is 69.1 out of 100. That is place 24 of 283 ranked models.
What is GPT-5.5 best at?
Knowledge: 70.4 out of 100, place 19 of 268.
How fast is GPT-5.5?
52 output tokens per second, 2.7 s to the first token.
What is the context window of GPT-5.5?
1.05M tokens.
Is GPT-5.5 open source?
No. Its weights are not published.
When was GPT-5.5 released?
April 2026.

Badge

GPT-5.5 on Inference360: place 24 of 283, index 69.1
HTML
<a href="https://inference360.ai/models/gpt-5-5"><img src="https://inference360.ai/badges/gpt-5-5.svg" alt="GPT-5.5 on Inference360: place 24 of 283, index 69.1" height="20"></a>
Markdown
[![GPT-5.5 on Inference360: place 24 of 283, index 69.1](https://inference360.ai/badges/gpt-5-5.svg)](https://inference360.ai/models/gpt-5-5)

Compare with

Grok 4.769.1MiniMax M3.1 Flash69.3GPT-5.6 Terra68.9DeepSeek V4.1 Flash 101768.5Gemini 3.7 Flash68.5

Lower price

Ling 3.1 Flash67.2 · $0.135/1MDeepSeek V4.1 Flash 101768.5 · $0.40/1MClaude Haiku 5.566.3 · $0.50/1M

More from OpenAI

All models
GPT-5.5 Pro74.8GPT Audio 1.5UnrankedGPT-5.4 Image 2UnrankedGPT Mini LatestUnrankedGPT Chat LatestUnranked