Kimi K2 Instruct by Moonshot AI is available through 79 API providers on LMSpeed. Compare API pricing from $0.0010 to $250.00 per million input tokens across providers. Free API access is offered by 8 providers. In speed benchmarks, the fastest provider reaches 227 tok/s.
Moonshot Kimi K2 Instruct is an instruction-tuned variant in the Kimi series, optimized for following instructions and conversational tasks.
Также известна как
Compare Kimi K2 Instruct API pricing across 71 providers. Input prices range from $0.0010 to $250.00 per million input. 3173721 API offers the lowest rate at $0.0010/M. 8 providers offer free API credits or a free tier.
| Провайдер | Вариант модели | Аудит | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен |
|---|---|---|---|---|---|---|
温云0% | moonshotai/kimi-k2-instruct-0905 | — | $75.00 | $75.00 | 66.4 t/s | 0.82 s |
| moonshotai/kimi-k2-instruct | — | $75.00 | $75.00 | — | — | |
老张API100% | kimi-k2-instruct | — | $0.537 | $0.537 | — | — |
柏拉图AI100% | moonshotai/kimi-k2-instruct | — | $0.548 | $2.19 | — | — |
星见雅 API100% | moonshotai/kimi-k2-instruct | — | Бесплатно | Бесплатно | — | — |
| 李/moonshotai/kimi-k2-instruct-0905 | — | Бесплатно | Бесплатно | — | — | |
| 英伟达/moonshotai/kimi-k2-instruct | — | Бесплатно | Бесплатно | — | — | |
| 英伟达/moonshotai/kimi-k2-instruct-0905 | — | Бесплатно | Бесплатно | — | — | |
MapleLeaf API100% | moonshotai/kimi-k2-instruct | — | $0.100 | $2.00 | — | — |
| moonshotai/kimi-k2-instruct-0905 | — | $1.00 | $3.00 | — | — | |
涵冰API99.9% | kimi-k2-instruct | — | $0.020 | $0.040 | — | — |
6345ywz API99.8% | moonshotai/kimi-k2-instruct | — | $0.137 | $0.411 | — | — |
AI Claw API93.2% | moonshotai/kimi-k2-instruct-0905 | — | $27.50 | $357.50 | — | — |
| moonshotai/kimi-k2-instruct | — | $250.00 | $6250.00 | — | — | |
Medu Chat100% | kimi-k2-Instruct-0905 | — | $0.0010 | $0.0010 | — | — |
素墨API100% | kimi-k2-instruct | — | $0.010 | $0.010 | — | — |
| kimi-k2-instruct-0905 | — | $0.010 | $0.010 | — | — | |
| kimi-k2-instruct-0905-tidy | — | $0.010 | $0.010 | — | — | |
| kimi-k2-instruct-tidy | — | $0.010 | $0.010 | — | — | |
| moonshotai/kimi-k2-instruct | — | $0.010 | $0.010 | — | — |
Данные о ценах из публичных API провайдеров
Сравните производительность по скорости и задержке у всех API-провайдеров.
| Провайдер | Скорость | Задержка | Тесты |
|---|---|---|---|
A3 kimi-k2-instruct-0905 | 226.90 tok/s | 0.29s | 10 |
moonshotai/kimi-k2-instruct-0905 | 182.40 tok/s | 1.38s | 70 |
moonshotai/kimi-k2-instruct | 78.08 tok/s | 0.52s | 20 |
kimi-k2-instruct-0905 | 74.54 tok/s | 0.63s | 5 |
moonshotai/kimi-k2-instruct-0905 | 66.42 tok/s | 0.82s | 10 |
moonshotai/Kimi-K2-Instruct:novita | 48.81 tok/s | 1.21s | 5 |
moonshotai/kimi-k2-instruct | 45.15 tok/s | 0.58s | 10 |
kimi-k2-instruct | 41.95 tok/s | 1.62s | 5 |
kimi-k2-instruct | 32.90 tok/s | 1.10s | 5 |
kimi-k2-instruct-0905 | 13.20 tok/s | 1.09s | 5 |
Показано 1–10 из 10 провайдеров
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.
gpt-oss
GPT-OSS is an open-weight language model family designed for self-hosted inference, research, and cost-efficient alternatives to proprietary GPT-class models.
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.
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.
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.
Рейтинги основаны на тестах, предоставленных сообществом, и периодических зондах работоспособности. Носит рекомендательный характер, не является официальными данными.
