Скорость вывода
5492.7 tok/s
5592.5 tok/s
56GPT-5.492.4 tok/s
5889.5 tok/s
GPT-5.4 by OpenAI is available through 381 API providers on LMSpeed. Compare API pricing from $0.0007 to $749250.00 per million input tokens across providers. Free API access is offered by 4 providers. Context window: 1,050,000. In speed benchmarks, the fastest provider reaches 428 tok/s.
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.
Также известна как
Compare GPT-5.4 API pricing across 377 providers. Input prices range from $0.0007 to $749250.00 per million input. XShuLab Sub2API offers the lowest rate at $0.0007/M. 4 providers offer free API credits or a free tier.
| Провайдер | Вариант модели | Аудит | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен |
|---|---|---|---|---|---|---|
6i2 API18.1% | gpt-5.4 | — | $0.100-96% | $0.290-98% | 428.4 t/s+363% | 4.17 s-97% |
331112 AI100% | gpt-5.4 | — | $2.50 | $15.00 | 86.2 t/s | 6.12 s-95% |
| gpt-5.4-openai-compact | — | $2.50 | $15.00 | — | — | |
TradingBase API100% | gpt-5.4 | — | $2.67 | $16.04 | 71.2 t/s | 1.63 s-99% |
星辰·AI100% | gpt-5.4 | — | $2.50 | $15.00 | 68.9 t/s | 7.53 s-94% |
| gpt-5.4-openai-compact | — | $2.50 | $15.00 | — | — | |
TokenClub API100% | gpt-5.4 | — | $2.50 | $15.00 | 68.0 t/s | 2.15 s-98% |
| gpt-5.4-openai-compact | — | $2.50 | $15.00 | — | — | |
VSLLM99.7% | gpt-5.4 | — | $0.036-99% | $0.036-100% | 59.9 t/s | 2.48 s-98% |
| gpt-5.4-openai-compact | — | $0.365-85% | $0.365-98% | — | — | |
KJK API99.6% | gpt-5.4 | — | $5.00 | $30.00 | 58.1 t/s | 4.45 s-97% |
QQ Code19.7% | gpt-5.4 | — | $0.385-85% | $2.31-85% | 55.0 t/s | 2.90 s-98% |
至强API0% | gpt-5.4 | — | $2.50 | $15.00 | 53.8 t/s | 2.45 s-98% |
Tokeness99.9% | gpt-5.4 | — | $1.63-35% | $9.75-35% | 53.2 t/s | 1.11 s-99% |
Claw API100% | gpt-5.4 | — | $0.342-86% | $2.05-86% | 53.1 t/s | 5.11 s-96% |
ABC Relay99.8% | gpt-5.4 | — | $0.342-86% | $2.05-86% | 52.6 t/s | 2.36 s-98% |
| gpt-5.4-openai-compact | — | $0.342-86% | $2.05-86% | — | — | |
Codex API100% | gpt-5.4 | — | $6.40 | $38.40 | 52.5 t/s | 3.21 s-98% |
| gpt-5.4-openai-compact | — | $6.40 | $38.40 | — | — | |
DNSHE0% | gpt-5.4 | — | Бесплатно | Бесплатно | 51.4 t/s | 5.46 s-96% |
Данные о ценах из публичных API провайдеров
Сравните производительность по скорости и задержке у всех API-провайдеров.
| Провайдер | Скорость | Задержка | Тесты |
|---|---|---|---|
6i2 API gpt-5.4 | 428.43 tok/s | 4.17s | 5 |
gpt-5.4 | 88.19 tok/s | 6.08s | 10 |
gpt-5.4 | 86.19 tok/s | 6.12s | 5 |
gpt-5.4(xhigh) | 85.16 tok/s | 5.13s | 5 |
gpt-5.4 | 71.20 tok/s | 1.63s | 5 |
gpt-5.4-fast | 69.82 tok/s | 2.96s | 5 |
gpt-5.4 | 68.88 tok/s | 7.53s | 5 |
gpt-5.4(high) | 68.13 tok/s | 7.10s | 5 |
gpt-5.4 | 68.03 tok/s | 2.15s | 5 |
gpt-5.4 | 62.33 tok/s | 4.41s | 5 |
gpt-5.4 | 62.20 tok/s | 3.35s | 15 |
gpt-5.4 | 61.54 tok/s | 4.56s | 5 |
gpt-5.4 | 61.34 tok/s | 3.48s | 10 |
gpt-5.4 | 60.78 tok/s | 5.82s | 5 |
gpt-5.4 | 59.85 tok/s | 2.48s | 5 |
gpt-5.4 | 59.26 tok/s | 8.63s | 15 |
gpt-5.4 | 58.12 tok/s | 4.03s | 5 |
gpt-5.4 | 58.08 tok/s | 4.45s | 25 |
gpt-5.4 | 56.80 tok/s | 2.27s | 10 |
gpt-5.4 | 56.64 tok/s | 4.52s | 5 |
Показано 1–20 из 108 провайдеров
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-5-2
OpenAI GPT-5.2 is a GPT-5 series model emphasizing advanced reasoning, multimodal understanding, and high-quality outputs for complex enterprise workloads.
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.
claude-haiku-4-5
Anthropic Claude Haiku 4.5 delivers fast, low-cost responses while retaining solid instruction following for chat, classification, and lightweight coding.
Рейтинги основаны на тестах, предоставленных сообществом, и периодических зондах работоспособности. Носит рекомендательный характер, не является официальными данными.Artificial Analysis
