Скорость вывода
7762.6 tok/s
7862.2 tok/s
80DeepSeek V4 Pro57.5 tok/s
8157.4 tok/s
DeepSeek V4 Pro by DeepSeek is available through 172 API providers on LMSpeed. Compare API pricing from $0.0007 to $17382.60 per million input tokens across providers. Free API access is offered by 14 providers. Context window: 1,048,576. In speed benchmarks, the fastest provider reaches 268 tok/s.
DeepSeek V4 Pro is the professional-tier DeepSeek V4 model, targeting frontier reasoning, coding, and agent workflows with maximum capability.
Также известна как
Compare DeepSeek V4 Pro API pricing across 158 providers. Input prices range from $0.0007 to $17382.60 per million input. VSLLM offers the lowest rate at $0.0007/M. 14 providers offer free API credits or a free tier.
| Провайдер | Вариант модели | Аудит | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен |
|---|---|---|---|---|---|---|
6i2 API18.1% | deepseek-v4-pro | — | $1.00 | $3.00 | 267.8 t/s+366% | 4.27 s |
Seamee API100% | deepseek-v4-pro | — | $1.74 | $3.48 | 80.6 t/s+40% | 5.17 s |
Tokeness99.9% | deepseek-v4-pro | — | $0.280-36% | $0.559-36% | 41.2 t/s | 9.83 s |
PICO API99.6% | deepseek-v4-pro | — | $0.411-6% | $0.822-6% | 39.5 t/s | 12.62 s |
CatClaw API100% | deepseek-v4-pro | — | $0.013-97% | $0.013-99% | 37.9 t/s | 6.37 s |
MyWebUI API84.3% | deepseek-v4-pro | — | $1.74 | $3.84 | 34.8 t/s | 26.35 s |
MKE AI100% | deepseek-v4-pro | — | $3.00 | $6.00 | 32.5 t/s | 7.44 s |
PICO AI47.4% | deepseek-v4-pro | — | $3.00 | $6.00 | 32.2 t/s | 11.32 s |
Hank Workspace API73.1% | deepseek-v4-pro | — | $75.00 | $75.00 | 30.9 t/s | 22.07 s |
| deepseek-v4-pro-nothinking | — | $75.00 | $75.00 | — | — | |
9527 API99.6% | deepseek-v4-pro | 70728084 | $3.00 | $6.00 | 27.0 t/s | 10.68 s |
| deepseek-ai/deepseek-v4-pro | — | $3.00 | $6.00 | — | — | |
星见雅 API100% | deepseek-v4-pro | — | Бесплатно | Бесплатно | 25.7 t/s | 14.88 s |
| Deepseek-V4-Pro | — | Бесплатно | Бесплатно | — | — | |
| deepseek-ai/deepseek-v4-pro | — | Бесплатно | Бесплатно | — | — | |
| 英伟达/deepseek-ai/deepseek-v4-pro | — | Бесплатно | Бесплатно | — | — | |
6345ywz API99.8% | deepseek-ai/deepseek-v4-pro | — | $2.05 | $2.05 | 13.8 t/s | 87.84 s |
| deepseek-ai/DeepSeek-V4-Pro | — | $0.274-37% | $0.822-6% | — | — | |
V-API100% | DeepSeek-V4-Pro | — | $0.080-82% | $0.080-91% | — | — |
| deepseek-v4-pro | — | $3.00 | $6.00 | — | — |
Данные о ценах из публичных API провайдеров
Сравните производительность по скорости и задержке у всех API-провайдеров.
| Провайдер | Скорость | Задержка | Тесты |
|---|---|---|---|
6i2 API deepseek-v4-pro | 267.79 tok/s | 4.27s | 10 |
deepseek-v4-pro | 80.55 tok/s | 5.17s | 5 |
deepseek-v4-pro | 48.13 tok/s | 4.11s | 5 |
deepseek-v4-pro | 41.24 tok/s | 9.83s | 15 |
deepseek/deepseek-v4-pro-free | 41.05 tok/s | 4.40s | 5 |
deepseek/deepseek-v4-pro | 40.48 tok/s | 5.82s | 5 |
deepseek-v4-pro | 40.04 tok/s | 5.19s | 5 |
deepseek-v4-pro | 39.85 tok/s | 15.07s | 5 |
deepseek-v4-pro | 39.54 tok/s | 12.62s | 5 |
deepseek-v4-pro | 37.90 tok/s | 6.37s | 5 |
deepseek-v4-pro | 37.68 tok/s | 5.78s | 130 |
deepseek-v4-pro | 37.67 tok/s | 7.34s | 5 |
deepseek/deepseek-v4-pro | 36.31 tok/s | 12.13s | 5 |
deepseek-v4-pro | 35.26 tok/s | 7.49s | 10 |
deepseek-v4-pro | 35.01 tok/s | 11.46s | 15 |
deepseek-v4-pro | 34.76 tok/s | 26.35s | 5 |
deepseek-v4-pro-nx | 33.42 tok/s | 5.04s | 5 |
deepseek-v4-pro | 32.55 tok/s | 7.44s | 5 |
deepseek-v4-pro | 32.26 tok/s | 15.36s | 5 |
deepseek-v4-pro | 32.24 tok/s | 11.32s | 5 |
Показано 1–20 из 27 провайдеров
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.
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.
glm-5-1
Zhipu GLM-5.1 is a next-generation GLM model aimed at frontier reasoning, coding, and bilingual agent applications.
kimi-k2-5
Moonshot Kimi K2.5 is an open-weight multimodal agent model with native vision and text input, strong coding performance, and a 256K context window.
deepseek-v3-2
DeepSeek V3.2 is an upgraded V3-series MoE model with stronger reasoning, coding, and math performance, widely available through OpenAI-compatible API relays.
minimax-m2-5
MiniMax M2.5 is MiniMax's flagship text model for coding and agents, with SOTA-level programming and agentic performance, improved token efficiency, and fast high-TPS API deployment.
Рейтинги основаны на тестах, предоставленных сообществом, и периодических зондах работоспособности. Носит рекомендательный характер, не является официальными данными.Artificial Analysis
