Скорость вывода
14190.4 tok/s
15185.7 tok/s
16Grok 4.20178.5 tok/s
18172.1 tok/s
Grok 4.20 is available through 84 API providers on LMSpeed. Compare API pricing from $0.0001 to $75.00 per million input tokens across providers. Free API access is offered by 7 providers. Context window: 2,000,000. In speed benchmarks, the fastest provider reaches 111 tok/s.
xAI Grok 4.20 is a Grok 4 series model with real-time knowledge integration, strong reasoning, and conversational capabilities for search-augmented chat.
Также известна как
Compare Grok 4.20 API pricing across 77 providers. Input prices range from $0.0001 to $75.00 per million input. Joverna offers the lowest rate at $0.0001/M. 7 providers offer free API credits or a free tier.
| Провайдер | Вариант модели | Аудит | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен |
|---|---|---|---|---|---|---|
星见雅 API100% | grok-4.20-0309 | — | Бесплатно | Бесплатно | 111.4 t/s | 1.33 s-90% |
AI Claw API93.2% | grok-4.20-fast | — | $50.00 | $250.00 | 92.3 t/s | 4.20 s-69% |
| grok-4.20-0309 | — | $75.00 | $1200.00 | — | — | |
6345ywz API99.8% | grok-4.20-fast | — | $0.274-86% | $0.548-91% | 90.1 t/s | 3.35 s-75% |
| grok-4.20-0309 | — | $0.274-86% | $1.37-77% | — | — | |
Yunchu API99.9% | grok-4.20-beta | — | $0.240-88% | $1.92-68% | 63.2 t/s | 0.79 s-94% |
| grok-4.20-fast | — | $0.800-60% | $0.800-87% | — | — | |
MineWuer API22% | grok-4.20-0309 | — | Бесплатно | Бесплатно | 62.0 t/s | 9.27 s-31% |
小水管 API100% | grok-4.20 | — | Бесплатно | Бесплатно | — | — |
MapleLeaf API100% | grok-4.20-0309 | — | $75.00 | $75.00 | — | — |
| grok-4.20-fast | — | $75.00 | $75.00 | — | — | |
Cuz AI100% | grok-4.20-0309 | — | $75.00 | $75.00 | — | — |
| grok-4.20-fast | — | $75.00 | $75.00 | — | — | |
AiXiaobai API100% | grok-4.20-fast | — | $17.47 | $17.47 | — | — |
Leonhard API99.8% | grok-4.20-fast | — | $0.0025-100% | $0.0025-100% | — | — |
RenRen API99.5% | grok-4.20-fast | — | $1.42-29% | $2.83-53% | — | — |
初叶🍂Furry API98.1% | grok-4.20-fast | — | $0.050-97% | $0.050-99% | — | — |
猫羽霖API76.2% | grok-4.20-fast | 70728088 | $0.050-97% | $0.050-99% | — | — |
Joverna100% | grok-4.20-fast | — | $0.0001-100% | $0.0001-100% | — | — |
素墨API100% | grok-4.20-beta | — | $0.010-99% | $0.010-100% | — | — |
Данные о ценах из публичных API провайдеров
Сравните производительность по скорости и задержке у всех API-провайдеров.
| Провайдер | Скорость | Задержка | Тесты |
|---|---|---|---|
星见雅 API grok-4.20-0309 | 111.36 tok/s | 1.33s | 5 |
grok-4.20-beta | 102.92 tok/s | 9.88s | 15 |
grok-4.20-fast | 92.30 tok/s | 4.20s | 5 |
grok-4.20-fast | 90.10 tok/s | 3.35s | 5 |
grok-4.20-beta | 77.51 tok/s | 11.28s | 5 |
grok-4.20 | 74.21 tok/s | 18.28s | 5 |
grok-4.20-beta | 63.19 tok/s | 0.79s | 5 |
grok-4.20-0309 | 61.96 tok/s | 9.27s | 5 |
grok-4.20-beta | 57.56 tok/s | 1.29s | 25 |
grok-4.20-beta | 54.81 tok/s | 1.17s | 5 |
grok-4.20-beta | 48.46 tok/s | 1.00s | 5 |
grok-4.20-beta | 40.76 tok/s | 1.13s | 5 |
Показано 1–12 из 12 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-5-2
OpenAI GPT-5.2 is a GPT-5 series model emphasizing advanced reasoning, multimodal understanding, and high-quality outputs for complex enterprise workloads.
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.
deepseek-v4-flash
DeepSeek V4 Flash is a fast, cost-efficient language model in the DeepSeek V4 family, optimized for low-latency chat, coding assistance, and high-throughput API workloads while retaining strong reasoning quality.
gemini-3-1-pro
Google Gemini 3.1 Pro is a Gemini 3 series model with advanced multimodal reasoning, long-context support, and strong performance on coding and analytical tasks.
Рейтинги основаны на тестах, предоставленных сообществом, и периодических зондах работоспособности. Носит рекомендательный характер, не является официальными данными.Artificial Analysis
