Скорость вывода
8353.9 tok/s
8453.3 tok/s
86Qwen3.551.8 tok/s
8751.7 tok/s
Qwen3.5 is available through 181 API providers on LMSpeed. Compare API pricing from $0.0049 to $5994.00 per million input tokens across providers. Free API access is offered by 17 providers. Context window: 262,144. In speed benchmarks, the fastest provider reaches 165 tok/s.
Alibaba Qwen3.5 is a Qwen3 generation model with improved reasoning, multilingual support, and efficient inference for chat, coding, and agent applications.
Также известна как
Compare Qwen3.5 API pricing across 164 providers. Input prices range from $0.0049 to $5994.00 per million input. ModelPool offers the lowest rate at $0.0049/M. 17 providers offer free API credits or a free tier.
| Провайдер | Вариант модели | Аудит | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен |
|---|---|---|---|---|---|---|
Real AI WAN81.7% | Qwen3.5-35B-A3B | — | $1.00 | $4.00 | 77.1 t/s+49% | 8.65 s |
| Qwen3.5-122B-A10B | — | $2.00 | $8.00 | 47.8 t/s | 20.41 s | |
| Qwen3.5-9B | — | $0.500-17% | $2.00-44% | — | — | |
| Qwen3.5-397B-A17B | — | $8.00 | $32.00 | — | — | |
星见雅 API100% | 英伟达/qwen/qwen3.5-397b-a17b | — | Бесплатно | Бесплатно | 21.3 t/s | 4.35 s |
| 李/qwen/qwen3.5-122b-a10b | — | Бесплатно | Бесплатно | — | — | |
| 李/qwen/qwen3.5-397b-a17b | — | Бесплатно | Бесплатно | — | — | |
| 英伟达/qwen/qwen3.5-122b-a10b | — | Бесплатно | Бесплатно | — | — | |
V-API100% | Qwen3.5-122B-A10B | — | $0.020-97% | $0.020-99% | — | — |
| Qwen3.5-35B-A3B | — | $0.020-97% | $0.020-99% | — | — | |
| qwen3.5-35b-a3b | — | $0.400-33% | $3.20-11% | — | — | |
| qwen3.5-27b | — | $0.600 | $4.80 | — | — | |
| qwen3.5-122b-a10b | — | $0.800 | $6.40 | — | — | |
| qwen3.5-397b-a17b | — | $1.20 | $7.20 | — | — | |
老张API100% | qwen3.5-35b-a3b | — | $0.058-90% | $0.460-87% | — | — |
| qwen3.5-27b | — | $0.086-86% | $0.690-81% | — | — | |
| qwen3.5-122b-a10b | — | $0.115-81% | $0.921-74% | — | — | |
| qwen3.5-397b-a17b | — | $0.173-71% | $1.04-71% | — | — | |
小水管 API100% | qwen3.5-122b-a10b | — | Бесплатно | Бесплатно | — | — |
| qwen3.5-27b | — | Бесплатно | Бесплатно | — | — |
Данные о ценах из публичных API провайдеров
Сравните производительность по скорости и задержке у всех API-провайдеров.
| Провайдер | Скорость | Задержка | Тесты |
|---|---|---|---|
Koru API qwen3.5-9b | 164.69 tok/s | 7.09s | 10 |
Qwen/Qwen3.5-9B | 108.96 tok/s | 7.99s | 5 |
Qwen/Qwen3.5-4B | 95.56 tok/s | 16.29s | 10 |
Qwen3.5-35B-A3B | 77.08 tok/s | 8.65s | 10 |
qwen3.5-27b | 63.56 tok/s | 13.78s | 5 |
qwen/qwen3.5-122b-a10b | 60.74 tok/s | 0.78s | 15 |
Qwen3.5-122B-A10B | 47.80 tok/s | 20.41s | 5 |
qwen/qwen3.5-122b-a10b | 44.50 tok/s | 2.97s | 90 |
qwen3.5-397b-a17b | 40.00 tok/s | 21.87s | 5 |
Qwen/Qwen3.5-27B | 34.21 tok/s | 33.63s | 5 |
Qwen/Qwen3.5-397B-A17B | 32.76 tok/s | 29.36s | 15 |
qwen/qwen3.5-397b-a17b | 29.51 tok/s | 8.50s | 90 |
英伟达/qwen/qwen3.5-397b-a17b | 21.31 tok/s | 4.35s | 5 |
qwen3.5 | 17.35 tok/s | 54.52s | 5 |
Показано 1–14 из 14 провайдеров
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-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.
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.
glm-5-1
Zhipu GLM-5.1 is a next-generation GLM model aimed at frontier reasoning, coding, and bilingual agent applications.
minimax-m2-7
MiniMax M2.7 is a high-tier M2-series model tuned for complex reasoning, long-context dialogue, and production-grade API workloads.
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.
Рейтинги основаны на тестах, предоставленных сообществом, и периодических зондах работоспособности. Носит рекомендательный характер, не является официальными данными.Artificial Analysis
