Скорость вывода
6976.3 tok/s
7075.5 tok/s
71GPT-5.272.8 tok/s
7271.0 tok/s
GPT-5.2 by OpenAI is available through 309 API providers on LMSpeed. Compare API pricing from $0.0005 to $17482.50 per million input tokens across providers. Free API access is offered by 1 provider. Context window: 128,000. In speed benchmarks, the fastest provider reaches 94 tok/s.
OpenAI GPT-5.2 is a GPT-5 series model emphasizing advanced reasoning, multimodal understanding, and high-quality outputs for complex enterprise workloads.
Также известна как
Compare GPT-5.2 API pricing across 308 providers. Input prices range from $0.0005 to $17482.50 per million input. XShuLab Sub2API offers the lowest rate at $0.0005/M. 1 provider offers free API credits or a free tier.
| Провайдер | Вариант модели | Аудит | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен |
|---|---|---|---|---|---|---|
Feng Love API99.8% | gpt-5.2 | — | $1.75 | $14.00 | 64.6 t/s | 4.77 s-96% |
BUZZ100% | gpt-5.2 | — | $1.74 | $13.95 | 64.2 t/s | 3.55 s-97% |
ModelPool97.7% | gpt-5.2 | — | $0.150-91% | $1.20-91% | 64.2 t/s | 5.26 s-95% |
| gpt-5.2-chat | — | $0.150-91% | $1.20-91% | — | — | |
| gpt-5.2-chat-latest | — | $0.150-91% | $1.20-91% | — | — | |
钠 API100% | gpt-5.2 | — | $0.041-98% | $0.329-98% | 60.0 t/s | 3.33 s-97% |
| gpt-5.2(auto) | — | $0.041-98% | $0.329-98% | — | — | |
| gpt-5.2(high) | — | $0.041-98% | $0.329-98% | — | — | |
| gpt-5.2(low) | — | $0.041-98% | $0.329-98% | — | — | |
| gpt-5.2(medium) | — | $0.041-98% | $0.329-98% | — | — | |
| gpt-5.2(xhigh) | — | $0.041-98% | $0.329-98% | — | — | |
WSocket AI99.3% | gpt-5.2 | — | $17482.50 | $139860.00 | 54.0 t/s | 1.45 s-99% |
Yun API99.9% | gpt-5.2-2025-12-11 | — | $0.240-86% | $1.92-86% | 32.4 t/s | 1.97 s-98% |
| gpt-5.2 | — | $0.0027-100% | $0.0027-100% | — | — | |
| gpt-5.2-chat-latest | — | $0.240-86% | $1.92-86% | — | — | |
| gpt-5.2-xhigh | — | $0.240-86% | $1.92-86% | — | — | |
| gpt-5.2-chat | — | $10.27 | $10.27-27% | — | — | |
| gpt-5.2-chat-2025-12-11 | — | $10.27 | $10.27-27% | — | — | |
柏拉图AI100% | gpt-5.2-all | — | $0.0068-100% | $0.0068-100% | — | — |
| gpt-5.2 | — | $0.240-86% | $1.92-86% | — | — |
Данные о ценах из публичных API провайдеров
Сравните производительность по скорости и задержке у всех API-провайдеров.
| Провайдер | Скорость | Задержка | Тесты |
|---|---|---|---|
PackyAPI gpt-5.2 | 93.71 tok/s | 11.49s | 10 |
gpt-5.2-chat | 71.48 tok/s | 2.89s | 5 |
gpt-5.2 | 64.56 tok/s | 4.77s | 5 |
gpt-5.2 | 64.23 tok/s | 3.55s | 5 |
gpt-5.2 | 64.21 tok/s | 5.26s | 5 |
gpt-5.2 | 64.17 tok/s | 2.95s | 5 |
gpt-5.2 | 60.05 tok/s | 3.33s | 5 |
gpt-5.2 | 58.71 tok/s | 1.63s | 10 |
gpt-5.2 | 57.02 tok/s | 4.60s | 5 |
gpt-5.2 | 56.63 tok/s | 1.90s | 5 |
gpt-5.2 | 55.45 tok/s | 1.43s | 5 |
gpt-5.2 | 53.95 tok/s | 1.45s | 5 |
gpt-5.2 | 53.46 tok/s | 1.88s | 10 |
gpt-5.2 | 52.28 tok/s | 2.67s | 15 |
gpt-5.2-chat | 51.41 tok/s | 2.14s | 5 |
gpt-5.2 | 50.67 tok/s | 2.72s | 5 |
gpt-5.2 | 48.12 tok/s | 2.93s | 10 |
gpt-5.2 | 48.10 tok/s | 1.54s | 5 |
gpt-5.2 | 46.23 tok/s | 8.11s | 5 |
GPT-5.2 | 45.37 tok/s | 1.95s | 5 |
Показано 1–20 из 23 провайдеров
gpt-5-4
OpenAI GPT-5.4 extends the GPT-5 family with stronger instruction following, deeper tool use, and improved performance on coding, math, and long-document analysis.
gpt-5-3-codex
OpenAI GPT-5.3 Codex is a code-specialized variant in the GPT-5 series, optimized for code generation, debugging, and software development tasks.
gpt-5-4-mini
OpenAI GPT-5.4 Mini is a compact language model in the GPT-5 series, optimized for quick responses and high throughput.
claude-opus-4-6
Anthropic Claude Opus 4.6 is the most capable Claude Opus tier, optimized for complex analysis, long-horizon coding, and high-stakes enterprise reasoning workloads.
claude-sonnet-4-6
Anthropic Claude Sonnet 4.6 extends the Sonnet line with improved tool use, coding reliability, and long-context performance for everyday production workloads.
gpt-5-2-codex
OpenAI GPT-5.2 Codex is a code-specialized variant in the GPT-5 series, optimized for code generation, debugging, and software development tasks.
Рейтинги основаны на тестах, предоставленных сообществом, и периодических зондах работоспособности. Носит рекомендательный характер, не является официальными данными.Artificial Analysis
