GPT-5.5 Pro

GPT-5.5 Pro sits within the gpt-pro family and is presented as a multimodal model that accepts text and image inputs while producing text outputs, with built-in support for attachments, reasoning, tool calling, and structured output. The combination of these capabilities suggests an intended focus on agentic workflows where the model must parse diverse input formats, plan multi-step solutions, and emit machine-readable responses for downstream systems. Practical deployments that involve parsing documents with embedded visuals, orchestrating external tools, or producing schema-constrained outputs align with this capability profile.

On third-party tracking, GPT-5.5 Pro appears on the llm-stats.com leaderboard where it is ranked 37th on the composite LLM Stats Score, with a Quality Tracker reading marked Stable at 0.00σ across seven-day, thirty-day, and all-time windows on a very small vote base. The same model profile includes a benchmark table where BrowseComp appears among the evaluations, a benchmark designed to test persistent web navigation and retrieval of hard-to-find entangled information. The limited independent evaluation footprint, paired with the model's wide input and context capacity, points to a system still building its public benchmark record while offering generous operating room for long-context reasoning tasks.

Neongpt-5-5-progpt-pro

Open provider page Provider docs

Quick Info

Provider: Neon
Model key: gpt-5-5-pro
Release date: Apr 23, 2026
Last updated: Apr 23, 2026
Knowledge cutoff: 2025-12-01
AI SDK package: @ai-sdk/openai
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $30.00
Output token cost: $180.00

Limits

Input tokens: 922,000 tokens
Output tokens: 128,000 tokens
Context window: 1,050,000 tokens

Transparent token rates

Compare GPT-5.5 Pro pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GPT-5.5 Pro

OpenCode Zen

Official sourceAnnouncement

OpenAI introduced GPT-5.5 and GPT-5.5 Pro on April 23, 2026, positioning the family as its "smartest and most intuitive" model yet, with particular gains in agentic coding, computer use, knowledge work, and early scientific research. GPT-5.5 Pro is rolling out to Pro, Business, and Enterprise users in ChatGPT, and the post was updated April 24, 2026, to confirm GPT-5.5 and GPT-5.5 Pro are also available in the API alongside an updated system card describing additional safeguards. The launch post emphasizes efficiency: GPT-5.5 matches GPT-5.4 per-token latency in real-world serving while operating at a higher intelligence level, and uses significantly fewer tokens to complete the same Codex tasks. OpenAI described its strongest set of safeguards to date, including evaluation across its safety and preparedness frameworks, internal and external red-teaming, targeted testing for advanced cybersecurity and biology capabilities, and feedback from nearly 200 trusted early-access partners before release.

Requesty

Coverage

Hidekazu Konishi's GPT Model Release Timeline (first published June 19, 2026; last updated September 16, 2026) compiles OpenAI's GPT lineage from GPT-1 in June 2018 through the GPT-5 generation in 2025 and 2026, mapping branching into multimodal "omni" models and the o-series reasoning track that converged into GPT-5. Within that lineage, GPT-5.5 Pro appears as part of the unified GPT-5 generation released in 2026, fitting the timeline's pattern of GPT-5.x iterative steps that build on GPT-5 and include dedicated reasoning-tuned Pro variants. The article serves as contextual lineage rather than breaking news for GPT-5.5 Pro specific

OpenCode Zen

CoverageBenchmark

LLM-Stats, a third-party benchmark tracker, places GPT-5.5 Pro at rank 39 on its composite LLM Stats Score as of October 6, 2026, with a Quality Tracker reading of 0.00σ over both 7-day and 30-day windows, indicating stable performance. The page lists per-benchmark rankings drawn from openai.com, including BrowseComp at rank 6 with 0.90/1, GDPval-MM at rank 2 with 0.82/1, Humanity's Last Exam at rank 11 with 0.57/1, and FrontierMath, giving independent visibility into Pro-tier evaluation. The tracker also displays how GPT-5.5 Pro holds up as conversation depth increases, providing a longitudinal view of capability at longer contexts, and reports a full ranked list of every benchmark the model has been evaluated on. Methodology notes indicate scores are sourced from OpenAI's scorecard, paper, or official blog posts, making this a useful corroboration layer for creator-reported performance data.

Requesty

CoverageBenchmark

Artificial Analysis lists GPT-5.5 Pro (xhigh) as an OpenAI proprietary reasoning model released April 2026, with text and image input, text output, and a 922k-token context window. The page is the "xhigh" reasoning-economy variant, and explicitly notes that a non-reasoning variant may also exist. The model's comparison Scraped Intelligence Index, Speed, and Cost fields show as Unknown or $0.00 because the relevant metrics are gated behind Artificial Analysis's premium login, so the excerpt should not be read as the model being free or unmeasured. Even without those numeric values, the page independently corroborates GPT-5.5 Pro's ide