llama-3-2-instruct
开发者: Meta
别名
Llama 3.2 Instruct（Meta）可通过 LMSpeed 上的 61 家 API 服务商获取。API 价格对比：输入价格从 $0.0001 到 $75.00/M，91VIP提供最低价。2 家服务商提供免费 API 额度。速度基准测试中，最快服务商达到 149 tok/s。
对比 Llama 3.2 Instruct 在 59 家服务商的 API 价格。输入价格从 $0.0001/M 到 $75.00/M，其中 91VIP 提供最低价 $0.0001/M。
| 服务商 | 模型变体 | 输入 ($/M) | 输出 ($/M) | 速度 (t/s) |
|---|---|---|---|---|
 OpenRouter Fans | meta-llama/llama-3.2-3b-instruct:free | Free | Free | — |
| MapleLeaf API | meta-llama/llama-3.2-3b-instruct:free | Free | Free | — |
| 91VIP | llama-3.2-1b-instruct | $0.0001 | $0.0000 | — |
| Futureppo | llama-3.2-1b-instruct | $0.0001 | $0.0000 | — |
| 91VIP | llama-3.2-3b-instruct | $0.0001 | $0.0000 | — |
| Futureppo | llama-3.2-3b-instruct | $0.0001 | $0.0000 | — |
 柏拉图AI | llama-3.2-1b-instruct | $0.0050 | $0.0050 | — |
 毫秒API | llama-3.2-1b-instruct | $0.0050 | $0.0050 | — |
| llama-3.2-1b-instruct | $0.0050 | $0.0050 | — | |
| 小豆包API | llama-3.2-1b-instruct | $0.0050 | $0.0050 | — |
 天絮 API | meta/llama-3.2-1b-instruct | $0.010 | $0.010 | — |
| Seamee API | meta/llama-3.2-1b-instruct | $0.010 | $0.010 | — |
| 素墨API | llama-3.2-1b-instruct | $0.010 | $0.010 | — |
| 素墨API | llama-3.2-3b-instruct | $0.010 | $0.010 | — |
| 素墨API | meta/llama-3.2-1b-instruct | $0.010 | $0.010 | — |
| 素墨API | meta/llama-3.2-3b-instruct | $0.010 | $0.010 | — |
| MapleLeaf API | meta/llama-3.2-1b-instruct | $0.010 | $0.010 | — |
| 天絮 API | meta/llama-3.2-3b-instruct | $0.020 | $0.020 | — |
| Seamee API | meta/llama-3.2-3b-instruct | $0.020 | $0.020 | — |
| MapleLeaf API | meta/llama-3.2-3b-instruct | $0.020 | $0.020 | — |
价格数据来自各服务商公开 API
Llama 3.2 Instruct 可通过 2 家服务商免费使用，无需按 token 付费。以下服务商提供免费 API 额度或免费套餐：
| 提供商 | 速度 |
|---|---|
| OpenRouter Fans | — |
| MapleLeaf API | — |
最新基准测试结果，测量 API 响应速度与首字延迟。
| 时间 | 模型 | 速度 | 延迟 |
|---|---|---|---|
| 02/18/2026, 08:39 | meta/llama-3.2-1b-instruct | 180.81 tok/s | 4.19s |
| 02/18/2026, 08:39 | meta/llama-3.2-1b-instruct | 211.52 tok/s | 5.69s |
| 02/18/2026, 08:39 | meta/llama-3.2-1b-instruct | 213.91 tok/s | 6.14s |
| 02/18/2026, 08:39 | meta/llama-3.2-1b-instruct | 14.86 tok/s | 5.37s |
| 02/18/2026, 08:39 | meta/llama-3.2-1b-instruct | 83.58 tok/s | 5.65s |
| 02/18/2026, 08:39 | meta/llama-3.2-3b-instruct | 110.23 tok/s | 0.97s |
| 02/18/2026, 08:39 | meta/llama-3.2-3b-instruct | 106.39 tok/s | 0.95s |
| 02/18/2026, 08:39 | meta/llama-3.2-3b-instruct | 96.91 tok/s | 0.45s |
| 02/18/2026, 08:39 | meta/llama-3.2-3b-instruct | 96.51 tok/s | 0.44s |
| 02/18/2026, 08:39 | meta/llama-3.2-3b-instruct | 96.95 tok/s | 0.45s |
| 02/18/2026, 08:27 | meta/llama-3.2-1b-instruct | 62.91 tok/s | 5.53s |
| 02/18/2026, 08:27 | meta/llama-3.2-1b-instruct | 227.93 tok/s | 5.55s |
| 02/18/2026, 08:27 | meta/llama-3.2-1b-instruct | 330.43 tok/s | 5.17s |
| 02/18/2026, 08:27 | meta/llama-3.2-1b-instruct | 16.67 tok/s | 4.74s |
| 02/18/2026, 08:27 | meta/llama-3.2-1b-instruct | 147.85 tok/s | 4.55s |
