GPT-5.2
Long-Horizon Agentic Work (Instant / Thinking / Pro)
GPT-5.2 pushed the GPT-5 line toward “do the work” tasks: long documents, multi-step plans, tool calling, and producing real artifacts like spreadsheets and presentations. It shipped as a series—Instant for speed, Thinking for depth, Pro for maximum quality—tightening the product pattern of one family covering multiple operating points.
Intelligence
Top 5%
Speed
Slow
66 tok/s output
Cost
High
$1.75 in / $14.00 out
Context
400K
Up to 400,000 tokens
How are Intelligence, Speed & Cost bucketed?
Intelligence and Speed buckets are percentile ranks on
Artificial Analysis. Cost buckets are fixed dollar
thresholds keyed off output-token price ($/M out).
Intelligence
- Top 1%≤ 1%
- Top 5%≤ 5%
- Top 10%≤ 10%
- Good≤ 25%
- Medium≤ 50%
- Below avg> 50%
Speed
- Top 1%≥ 345 tok/s
- Top 5%≥ 237 tok/s
- Top 10%≥ 196 tok/s
- Good≥ 146 tok/s
- Medium≥ 90 tok/s
- Slow< 90 tok/s
Cost
- Freeopen weights · self-host
- Low< $1 / M out
- Moderate$1–5 / M out
- High≥ $5 / M out
Why it matters
It cemented the modern UX: one model family with multiple tiers (Instant/Thinking/Pro) and rapid generation updates.
Core Capabilities
Long Documents
Handles entire codebases, books, and multi-doc RAG.
Multimodal
Combines text, vision, and audio in one model.
Generative
Produces images, video, audio, or other media.
Reasoning
Solves complex math, logic, and planning tasks.
Context Window
400k tokens
≈ 308+ pages
4k Chat 聊天
32k Long docs 长文档
128k Books 整本书
400k This model 本模型
1M Codebase 整个代码库
10M
Availability
API
Available
Product / App
Not available
Open Source
Not released
Enterprise
Contact sales
Pricing Model
Pay per token
Input and output billed separately.
Pay-per-token Capability / Performance
Where this model sits relative to the middle 60% of models in the tree. All scores are 0–10 (higher is better).
Lower 20% Upper 80% This model
Context / memory
Context window size · log-scaled
6.0
9.0
7.7
Lower 20% 20th percentile — 20% of models score below this This model Where the current model lands Upper 80% 80th percentile — only 20% of models score above this
Percentile boundaries are computed across every model in the tree that reports the underlying benchmark for each capability.
What it feels like
- Pro variant hits 93.2% on GPQA Diamond and 40.3% on FrontierMath Tier 1-3 — new state-of-the-art on hard math/science
- 400K-token context window — biggest OpenAI window yet, designed for long-running agents
- GDPval state-of-the-art — outperforms industry professionals on well-specified tasks across 44 occupations
- 55.6% on SWE-Bench Pro and 80% on SWE-bench Verified — highest-ever OpenAI coding scores
- Mixed reception: benchmark-strong but Reddit/X users call Instant 'bland' and 'corporate compliance training'-ish
- 40% price increase over GPT-5.1 — OpenAI's own system card acknowledges 'regressions in certain modes'
Best use cases
- Hardest math/science research where Pro/Thinking modes pull ahead of every competitor
- Long-running professional agent workflows (400K context + GDPval-leading task quality)
- Knowledge work spanning 44 occupations — spreadsheets, presentations, multi-step projects
- Coding agents at the highest tier where SWE-bench Pro 55.6% matters
Tools to try
Not ideal for
- Casual chat or creative writing — Instant tier feels muted compared to GPT-5.1
- Cost-sensitive bulk inference — 40% price hike vs 5.1 with no automatic upgrade win
- Workflows where GPT-5.1's adaptive thinking and faster output mattered more than peak IQ
Model Evolution
GPT is OpenAI's language model family.