GPT-5.2
openai/gpt-5.2- OpenAI
- Family · GPT-5
- Released 11 Dec 2025
- File · Image · Text → Text
- Reasoning · adjustable
- Tools · JSON · Vision
Spec & list price
Context391K
Max output128K
Input / 1M$1.75
Output / 1M$14.008× input
Scores · standing
Intelligence30.4#63/489
Usage
Tokens / day3.3B▼ 13% 7d
Usage rank#139
Requests / day167.5K
Tool calls / req0.22
Per request · 30 days
Avg input7.0K0.5× median
Avg output3090.4× median
Cache hit68%about median
Reasoning share50%0.9× median
Tool-call failures0.3%0.2× median
Consumption
Peak 16.4B 27 OctLow 1.8B 31 OctRank today #139
News & trends
2 mentions this week ▼ 86%this model · videos, posts, articles · 30 days
Trending · 14 days · GPT-5 family
- GPT-5.6 Sol can beat Factorio on peaceful mode175 points
- The Current Models: GPT-5.5 & GPT-5.6 (Sol, Terra, Luna) | ChatGPT 2026 (Ep. 4)7.8K views
- @ArtificialAnlys on X43.1K views
- OpenAI researcher defects to Anthropic after shipping GPT-51 upvotes
- [OC] I checked 163M Reddit posts and comments. Em dashes took off after GPT-4o, then fell 17% the month GPT-5 came out403 upvotes
- Codex 逆袭：从失败到 ChatGPT 底座 | OpenAI | ChatGPT | GPT-5 | Claude Code | o1 | 思维链 | 创新者的窘境 | Sam Altman28.1K views
News · latest from OpenAI
- Chinese labs now target OpenAI's GPT-6 Astra for distillationfork.zone · 1 Nov
- GPT-6.1 Terrasimonwillison.net · 31 Oct
- GPT-5.6 Sol can beat Factorio on peaceful moderyanmadden.net · 31 Oct
- Post-training 4B model with RL to fly Tesla Roadster 21% better than GPT-6 Astraskypilot.ai · 30 Oct
Benchmarks
bar = share of the top score · Benchmark data: Artificial AnalysisSpeed · median across 3 providers, latest day
Output speed#219/33352 tok/s1.3× slower than median
Fastest provider90 tok/sbest of 3 providers
First token#2542.06 s2.0× longer than median
#5 cheapest of 10 with a similar score
Value
Intelligence Index
Blended price per 1M tokens (log)
Family progress
Intelligence Index
- Family releases38.4
Technical details
- Reasoning
- on · medium
- Tokenizer
- GPT
- Moderation
- moderated
Reasoning effort
- xhigh
- high
- medium · default
- low
- none
Request parameters 11
- tools
- tool_choice
- response_format
- structured_outputs
- reasoning
- reasoning_effort
- include_reasoning
- max_tokens
- max_completion_tokens
- seed
+1 more
- verbosity
Interesting facts
- 2× the output room of similar models (128.0K against a 65.5K median).
- Weekdays average 7.5B tokens a day against 3.6B on weekends.
When to use
Best for
- Math and competition-style reasoning
- Coding with deep reasoning effort
- Long document analysis
- Image and file understanding
- Offline bulk jobs via the batch variant
Think twice for
- Low-cost high-volume chat
- Self-hosting (closed weights)
- Latency-critical interactive apps
Research papers
- CuraLoc: Repository-Level Fault Localization via Planned Code Inspection22 Oct
- UniGuard: Multi-View Guardrail Learning for Cross-Interface Safety Generalization22 Oct
- The Butterfly Effect in LLM Social Simulations: From Architecture Sensitivity to Robustness Audits20 Oct
- Same Simulation, Different Evidence: Where Retrieved Tool Output Costs LLM Agents Accuracy19 Oct
- Budgeted Attack Scheduling for LLM Red-Teaming: The Roles of Static Priors, Semantic Guidance, and Online Feedback19 Oct
- How Well Can LLMs Verbalize Mathematical Formulas? A Multi-Domain Evaluation17 Oct
- Beyond Generation: Using Multimodal LLMs as Detectors of LLM-Generated Deepfake Facial Images15 Oct
- Reinforcement Learning In Partially Verifiable Environments for Colon Cancer Treatment Planning12 Oct
Social
Most watched and discussed · 90 days
- GPT-5.2 pasó de 0 % a 75 % de victorias con una fecha límite #ShortsSTRATEGA · 67 views · 20 Oct
- Code Red bei OpenAI: GPT 5.2 soll Gemini und Claude schlagen!TL;DW deutschland · 52 views · 21 Oct
- GPT-5.2 vs GPT-5 mini: the river-crossing riddle test #AI103 #shortsLoop Learning · 38 views · 30 Oct
- GPT-5.2: stronger math #ShortsDigitalWeekly-Ai · 34 views · 23 Oct
- ChatGPT का इतिहास | 2017 से GPT-5.2 तक का सफर | TheBookStoryAudioThe Book Story · 24 views · 26 Oct
- GPT-5.2 và Gemini 3 Flash Thống Trị Bảng Xếp Hạng Toán Học AIAI Race - Cuộc Đua AI · 16 views · 20 Oct
- GPT-5.2 vs Gemini 3: The Test That Actually MattersSallØ Spartan 907 · 9 views · 22 Oct
- #31 GPT 5.2, Security meltdown in React, Cursor visual editor, Figma becomes Photoshop killerCmd+AI · 8 views · 20 Oct
- GPT-5.2 lập cú sốc toán học, cướp ngôi đầu của Google ngay sát nútAI Race - Cuộc Đua AI · 3 views · 21 Oct
- GPT-5.2 Advances Math & Science — OpenAI’s DOE, BBVA, Academy, and Monitorability ResearchDigitalWeekly-Ai · 2 views · 23 Oct
Same family
GPT-5- 23.0
- 38.4
- 39.0
- 24.7
Alternatives
other vendors · similar capability- 30.1
- 29.7
- 34.0
- 33.6
Using GPT-5.2 at volume?
Sign in to ask for a volume discount. You will continue in the OSMEXO dashboard.