GPT-5.5: OpenAI's Flagship From April to July 2026

What GPT-5.5 is

GPT-5.5 was OpenAI's flagship model between April and July 2026, its system card published on 23 April. It carries a 1,050,000 token context window and returns up to 128,000 output tokens, priced at $5 per million input tokens and $30 per million output. It was superseded on 9 July 2026 by the GPT-5.6 family — Luna, Terra and Sol —. The model remains available through OpenAI's API under the alias gpt-5.5, and OpenAI publish separate model guidance for it.

Regional processing uplift

The standard rate is $5 per million input tokens and $30 output. Regional processing endpoints, the data-residency options, are charged a 10% uplift on GPT-5.5.

OpenAI’s comparison with GPT-5.6

OpenAI's GPT-5.6 announcement uses GPT-5.5 as its baseline. On ExploitBench 2, which tracks progress from reaching vulnerable code to arbitrary code execution, GPT-5.6 Sol scores 73.5% against GPT-5.5's 47.9% at a comparable output-token budget. On SEC-Bench Pro, 71.2% against 45.8%. On ExploitGym, which asks agents to turn real vulnerabilities into working exploits, Sol nearly doubles 5.5's peak pass rate, from 15.1% to 24.9% under a two-hour cap. Away from security, OpenAI state that GPT-5.6 Luna — the smallest of the three — nearly matches GPT-5.5's peak performance at less than half the estimated cost, and that Terra surpasses it at lower cost. A customer, Rogo, reported a 6.2-point rubric-quality gain and 3.6-point accuracy gain over GPT-5.5 on their finance benchmark. These are vendor figures.

LMArena

On LMArena's text leaderboard GPT 5.5 has 1477 Elo (rank 32); by category, 1510 Elo in Coding (rank 59), 1500 Elo in Math (rank 19), 1510 Elo in Expert (rank 31) and 1448 Elo in Creative Writing (rank 42). An independently recorded deep-research score of 54 rounds out the picture.

API details

OpenAI's model page lists GPT-5.5 with reasoning effort settings of none, low, medium (the default), high and xhigh, and a knowledge cutoff of 1 December 2025. Cached input is $0.50 per million tokens. Prompts with more than 272,000 input tokens are priced at twice the input rate and 1.5 times the output rate for the full session, on standard, batch and flex processing. The alias gpt-5.5 points to the snapshot gpt-5.5-2026-04-23. The model is available on the Responses and Chat Completions endpoints, the Realtime and Live endpoints, the Batch API and the Assistants API, among others listed on its page; tool-specific models such as search and computer use carry a fee per tool call.

Rate limits

OpenAI publish rate limits by usage tier. Tier 1 allows 500 requests and 500,000 tokens per minute; Tier 2 5,000 requests and 1,000,000 tokens; Tier 3 5,000 requests and 2,000,000 tokens; Tier 4 10,000 requests and 4,000,000 tokens; and Tier 5 15,000 requests and 40,000,000 tokens per minute. Batch queue limits rise from 1,500,000 tokens at Tier 1 to 15 billion at Tier 5. The tier rises automatically with API usage and spend. For comparison, OpenAI's page lists GPT-5.4 at $2.50 per million input tokens and GPT-5.4 Mini at $0.75.

Official Resources