Choose a model to open its comparison page.
Llama 3.3 70B Instruct API pricing covers 27 API providers, from $0.049/M to $375.00/M. Llama 3.3 70B Instruct free API options are available from 1 provider.
The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model...
Input and output token limits for this model, plus how it ranks on long-context understanding.
Compare Llama 3.3 70B Instruct API pricing across 26 providers. Prices range from $0.049/M to $375.00/M. 云AI offers the lowest rate at $0.049/M. 1 provider offers free API credits or a free tier.
| Провайдер | Работоспособность | Вариант модели | Группа | Входные данные ($/M) | Выходные данные ($/M) | Скорость (т/с) | Первый токен | Аудит |
|---|---|---|---|---|---|---|---|---|
L1 100% | llama-3.3-70b-instruct | enterprise_emergency_standby | $57.53/M | $57.53/M | — | — | — | |
L1 100% | llama-3.3-70b-instruct | 特供-az 0.7 | $0.128/M | $0.128/M | — | — | — | |
L1 100% | llama-3.3-70b-instruct | 特供-az 0.7 | $0.128/M | $0.128/M | — | — | — | |
Dext API Бесплатно | L1 100% | llama-3.3-70b-instruct | 公益 | Бесплатно | Бесплатно | — | — | — |
L1 99% | meta/llama-3.3-70b-instruct | default | $375.00/M | $375.00/M | — | — | — | |
L1 96% | llama-3.3-70b-instruct | 小叶币 | $0.050/request | - | — | — | — | |
L1 100% | llama-3.3-70b-instruct | 特供-az 0.7 | $0.069/M | $0.069/M | — | — | — | |
L1 0% | llama-3.3-70b-instruct | Model | $0.049/M | $0.049/M | — | — | — |
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.
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-haiku-4-5
Anthropic Claude Haiku 4.5 delivers fast, low-cost responses while retaining solid instruction following for chat, classification, and lightweight coding.
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.
