llama-4-maverick-128e-instruct
开发者: Meta
别名
Llama 4 Maverick 128e Instruct（Meta）可通过 LMSpeed 上的 21 家 API 服务商获取。API 价格对比：输入价格从 $0.0005 到 $75.00/M，91VIP提供最低价。1 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 826 tok/s。
对比 Llama 4 Maverick 128e Instruct 在 20 家服务商的 API 价格。输入价格从 $0.0005/M 到 $75.00/M，其中 91VIP 提供最低价 $0.0005/M。
| 服务商 | 模型变体 | 输入 ($/M) | 输出 ($/M) | 速度 (t/s) |
|---|---|---|---|---|
| MapleLeaf API | meta/llama-4-maverick-17b-128e-instruct | Free | Free | — |
| 91VIP | llama-4-maverick-17b-128e-instruct | $0.0005 | $0.0008 | — |
| Futureppo | llama-4-maverick-17b-128e-instruct | $0.0005 | $0.0008 | — |
| 素墨API | llama-4-maverick-17b-128e-instruct | $0.010 | $0.010 | — |
| 素墨API | meta/llama-4-maverick-17b-128e-instruct | $0.010 | $0.010 | — |
 星见雅 API | 英伟达/meta/llama-4-maverick-17b-128e-instruct | $0.100 | $0.100 | — |
| 对空六课 API | 英伟达/meta/llama-4-maverick-17b-128e-instruct | $0.100 | $0.100 | — |
| WSocket AI | meta/llama-4-maverick-17b-128e-instruct | $0.200 | $0.200 | — |
| 人人 API | llama-4-maverick-17b-128e-instruct | $0.300 | $0.500 | — |
| SWT-API | llama-4-maverick-17b-128e-instruct | $0.375 | $0.375 | — |
| Synapse | Llama-4-Maverick-17B-128E-Instruct | $1.00 | $1.00 | — |
| 钱多多 API | meta/llama-4-maverick-17b-128e-instruct | $2.00 | $2.00 | — |
| Seamee API | LLM-Research/Llama-4-Maverick-17B-128E-Instruct | $2.00 | $2.00 | — |
 天絮 API | llama-4-maverick-17b-128e-instruct | $75.00 | $75.00 | — |
| 天絮 API | meta/llama-4-maverick-17b-128e-instruct | $75.00 | $75.00 | — |
 IXIOCCAPI | llama-4-maverick-17b-128e-instruct | $75.00 | $75.00 | — |
| MIX API | meta/llama-4-maverick-17b-128e-instruct | $75.00 | $75.00 | — |
| Seamee API | Meta/Llama-4-Maverick-17B-128E-Instruct | $75.00 | $75.00 | — |
| Seamee API | meta/llama-4-maverick-17b-128e-instruct | $75.00 | $75.00 | — |
| 温云 | meta/llama-4-maverick-17b-128e-instruct | $75.00 | $75.00 | — |
价格数据来自各服务商公开 API
Llama 4 Maverick 128e Instruct 可通过 1 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 |
|---|---|
| MapleLeaf API | — |
| 提供商 | 速度 | 延迟 | 测试次数 |
|---|---|---|---|
 GPT Load (Shiho)llama-4-maverick-17b-128e-instruct | 825.70 tok/s | 0.41s | 5 |
 星见雅 APImeta/llama-4-maverick-17b-128e-instruct | 131.83 tok/s | 0.48s | 5 |
 meta/llama-4-maverick-17b-128e-instruct | 100.41 tok/s | 0.21s | 15 |
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 |
|---|---|---|---|
| 03/31/2026, 10:16 | meta/llama-4-maverick-17b-128e-instruct | 107.02 tok/s | 0.32s |
| 03/31/2026, 10:16 | meta/llama-4-maverick-17b-128e-instruct | 98.43 tok/s | 0.19s |
| 03/31/2026, 10:16 | meta/llama-4-maverick-17b-128e-instruct | 93.75 tok/s | 0.18s |
| 03/31/2026, 10:16 | meta/llama-4-maverick-17b-128e-instruct | 106.97 tok/s | 0.17s |
| 03/31/2026, 10:16 | meta/llama-4-maverick-17b-128e-instruct | 104.39 tok/s | 0.17s |
| 03/31/2026, 10:15 | meta/llama-4-maverick-17b-128e-instruct | 104.12 tok/s | 0.29s |
| 03/31/2026, 10:15 | meta/llama-4-maverick-17b-128e-instruct | 96.01 tok/s | 0.17s |
| 03/31/2026, 10:15 | meta/llama-4-maverick-17b-128e-instruct | 93.31 tok/s | 0.17s |
| 03/31/2026, 10:15 | meta/llama-4-maverick-17b-128e-instruct | 101.97 tok/s | 0.22s |
| 03/31/2026, 10:15 | meta/llama-4-maverick-17b-128e-instruct | 105.10 tok/s | 0.17s |
| 03/31/2026, 10:14 | meta/llama-4-maverick-17b-128e-instruct | 98.62 tok/s | 0.30s |
| 03/31/2026, 10:14 | meta/llama-4-maverick-17b-128e-instruct | 98.75 tok/s | 0.21s |
| 03/31/2026, 10:14 | meta/llama-4-maverick-17b-128e-instruct | 89.16 tok/s | 0.19s |
| 03/31/2026, 10:14 | meta/llama-4-maverick-17b-128e-instruct | 108.21 tok/s | 0.20s |
| 03/31/2026, 10:14 | meta/llama-4-maverick-17b-128e-instruct | 100.29 tok/s | 0.19s |
| 03/13/2026, 21:30 | meta/llama-4-maverick-17b-128e-instruct | 140.78 tok/s | 0.56s |
| 03/13/2026, 21:30 | meta/llama-4-maverick-17b-128e-instruct | 125.11 tok/s | 0.56s |
| 03/13/2026, 21:30 | meta/llama-4-maverick-17b-128e-instruct | 124.89 tok/s | 0.46s |
| 03/13/2026, 21:30 | meta/llama-4-maverick-17b-128e-instruct | 124.47 tok/s | 0.40s |
| 03/13/2026, 21:30 | meta/llama-4-maverick-17b-128e-instruct | 143.88 tok/s | 0.39s |
