qwen2-5-instruct
Разработчик: Alibaba
Alibaba Qwen2.5 Instruct is an instruction-tuned language model in the Qwen2.5 series, optimized for chat, coding, and general assistant workloads.
Также известна как
Qwen2.5 Instruct by Alibaba is available through 107 API providers on LMSpeed. Compare API pricing from $0.0061 to $535.71 per million input tokens across providers. Free API access is offered by 1 provider. In speed benchmarks, the fastest provider reaches 93 tok/s.
Compare Qwen2.5 Instruct API pricing across 106 providers. Input prices range from $0.0061 to $535.71 per million input. ModelPool offers the lowest rate at $0.0061/M. 1 provider offers free API credits or a free tier.
| Провайдер | Вариант модели | Аудит | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен |
|---|---|---|---|---|---|---|
V-API100% | Qwen2.5-32B-Instruct | — | $0.015 | $0.015 | — | — |
| Qwen2.5-72B-Instruct | — | $0.020 | $0.020 | — | — | |
| qwen2.5-14b-instruct | — | $1.00 | $3.00 | — | — | |
| qwen2.5-32b-instruct | — | $2.00 | $6.00 | — | — | |
| qwen2.5-72b-instruct | — | $4.00 | $12.00 | — | — | |
柏拉图AI100% | qwen2.5-3b-instruct | — | $0.041 | $0.123 | — | — |
| qwen2.5-7b-instruct | — | $0.068 | $0.137 | — | — | |
| qwen2.5-14b-instruct | — | $0.137 | $0.411 | — | — | |
| qwen2.5-32b-instruct | — | $0.274 | $0.822 | — | — | |
| qwen2.5-72b-instruct | — | $0.548 | $1.64 | — | — | |
| Qwen/Qwen2.5-7B-Instruct | — | $10.27 | $10.27 | — | — | |
天絮 API100% | Qwen/Qwen2.5-7B-Instruct | — | $0.357 | $0.357 | — | — |
| Qwen/Qwen2.5-14B-Instruct | — | $0.714 | $0.714 | — | — | |
| Qwen/Qwen2.5-32B-Instruct | — | $1.29 | $1.29 | — | — | |
| Qwen/Qwen2.5-72B-Instruct | — | $4.21 | $4.21 | — | — | |
| Qwen/Qwen2.5-72B-Instruct-128K | — | $4.21 | $4.21 | — | — | |
| LoRA/Qwen/Qwen2.5-14B-Instruct | — | $535.71 | $535.71 | — | — | |
| LoRA/Qwen/Qwen2.5-32B-Instruct | — | $535.71 | $535.71 | — | — | |
| LoRA/Qwen/Qwen2.5-72B-Instruct | — | $535.71 | $535.71 | — | — | |
| LoRA/Qwen/Qwen2.5-7B-Instruct | — | $535.71 | $535.71 | — | — |
Данные о ценах из публичных API провайдеров
Сравните производительность по скорости и задержке у всех API-провайдеров.
| Провайдер | Скорость | Задержка | Тесты |
|---|---|---|---|
丸美小沐 qwen2.5-72b-instruct | 92.64 tok/s | 2.57s | 10 |
Qwen/Qwen2.5-14B-Instruct | 71.24 tok/s | 0.65s | 10 |
LoRA/Qwen/Qwen2.5-32B-Instruct | 64.68 tok/s | 0.84s | 5 |
Qwen/Qwen2.5-32B-Instruct | 58.60 tok/s | 0.73s | 20 |
LoRA/Qwen/Qwen2.5-14B-Instruct | 58.12 tok/s | 1.02s | 5 |
qwen2.5-14b-instruct | 49.95 tok/s | 0.96s | 5 |
qwen2.5-instruct | 48.51 tok/s | 0.58s | 15 |
Qwen2.5-7B-Instruct | 44.65 tok/s | 0.50s | 5 |
Qwen/Qwen2.5-7B-Instruct | 43.96 tok/s | 0.91s | 5 |
qwen2.5-7b-instruct | 42.04 tok/s | 1.36s | 15 |
Qwen/Qwen2.5-7B-Instruct | 39.38 tok/s | 1.40s | 35 |
Qwen2.5-32B-Instruct | 36.66 tok/s | 0.58s | 55 |
qwen2.5-7b-instruct | 34.15 tok/s | 1.01s | 5 |
Qwen/Qwen2.5-72B-Instruct | 32.52 tok/s | 0.79s | 10 |
Qwen/Qwen2.5-72B-Instruct | 29.84 tok/s | 1.01s | 40 |
qwen2.5-72b-instruct | 28.17 tok/s | 0.97s | 5 |
Vendor-A/Qwen/Qwen2.5-72B-Instruct | 27.74 tok/s | 1.25s | 5 |
LoRA/Qwen/Qwen2.5-72B-Instruct | 24.64 tok/s | 3.37s | 5 |
qwen2.5-72b-instruct | 22.88 tok/s | 1.08s | 5 |
Pro/Qwen/Qwen2.5-7B-Instruct | 20.74 tok/s | 0.87s | 5 |
Показано 1–20 из 22 провайдеров
deepseek-v3-2
DeepSeek V3.2 is a large language model in the DeepSeek V3 series, offering advanced reasoning, code generation, and multimodal capabilities.
kimi-k2-5
Moonshot Kimi K2.5 is a large language model in the Kimi series, offering advanced reasoning, code generation, and multimodal capabilities.
deepseek-v3
DeepSeek V3 is DeepSeek flagship MoE language model with 671B total parameters, delivering strong performance in reasoning, coding, and multilingual tasks at competitive inference cost.
qwen3
Alibaba Qwen3 is the Qwen family's flagship LLM series with dense and MoE variants, seamless thinking/non-thinking modes, and leading open-source performance in math, code, and agent tasks.
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
Zhipu GLM-5 is Zhipu flagship GLM series model with enhanced reasoning, agent capabilities, and strong performance on Chinese enterprise and coding scenarios.
Рейтинги основаны на тестах, предоставленных сообществом, и периодических зондах работоспособности. Носит рекомендательный характер, не является официальными данными.
