openai

GPT-5.5

gpt

Default frontier GPT for coding, computer use, research, and knowledge work

ReasoningTool CallingAttachmentsStructured Output

Context Window

1.1M

Max Output

128K

Temperature

No

Open Weights

No

Knowledge Cutoff

2025-12-01

Released

2026-04-23

Last Updated

2026-04-23

Modalities

Input:TextImagePDF
→
Output:Text

Benchmarks

BenchmarkScoreMetricHarness
SWE-Bench Pro58.6resolve rate-
Terminal-Bench78.2success rateTerminus-2
SWE-Atlas Codebase QnA45.43scoreCodex
SWE-Atlas Refactoring44.79scoreCodex
SWE-Atlas Test Writing42.59scoreCodex
Artificial Analysis Coding Agent Index65.3average pass@1Codex
SWE-Atlas Codebase QnA80.8pass@1Codex
SWE-Bench Pro30.9pass@1Codex
Terminal-Bench84.1pass@1Codex
Artificial Analysis Coding Agent Index60.4average pass@1Codex
SWE-Atlas Codebase QnA79.1pass@1Codex
SWE-Bench Pro26.2pass@1Codex
Terminal-Bench75.8pass@1Codex
Artificial Analysis Coding Agent Index57.8average pass@1Cursor CLI
SWE-Atlas Codebase QnA75pass@1Cursor CLI
SWE-Bench Pro24.9pass@1Cursor CLI
Terminal-Bench73.4pass@1Cursor CLI

Available Providers (13)

ProviderInput /1MOutput /1MCache Read /1MCache Write /1MReasoningStatus
Vercel AI GatewayVercel AI GatewayRp 90.169Rp 541.012Rp 9.017—effort—
ZenMuxZenMuxRp 90.169Rp 541.012Rp 9.017—effortExperimental
Merge GatewayMerge GatewayRp 90.169Rp 541.012Rp 9.017—effortExperimental
OrcaRouterOrcaRouterRp 90.169Rp 541.012Rp 9.017—✓Experimental
NEAR AI CloudNEAR AI CloudRp 90.169Rp 541.012Rp 9.017—effort—
Cloudflare AI GatewayCloudflare AI GatewayRp 90.169Rp 541.012Rp 9.017—effort—
PoePoeRp 81.973Rp 491.828Rp 8.197—effort—
Perplexity AgentPerplexity AgentRp 90.169Rp 541.012Rp 9.017—effort—
GMI CloudGMI CloudRp 90.169Rp 541.012Rp 9.017—effort—
Kilo GatewayKilo GatewayRp 90.169Rp 541.012Rp 9.017—effort—
OpenRouterOpenRouterRp 90.169Rp 541.012Rp 9.017—effort—
FastRouterFastRouterRp 90.169Rp 541.012——effort—
NanoGPTNanoGPTRp 90.169Rp 541.012Rp 9.017—effort—
← Kembali ke Katalog