Gemma 4 31B
google/gemma-4-31b-it
输入
$0.20
每 1M token
输出
$0.40
每 1M token
缓存输入
—
每 1M token
另有免费但限流的版本可用。
Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output. Features a 256K token context window, configurable thinking/reasoning mode, native function...
以下描述来自 OpenRouter，原文为英文。
- 发布时间
- 上下文窗口
- 262K token
- 最大输出
- 16K token
- 输入
- 图像, 文本, 视频
- 输出
- 文本
- 推理能力
- 支持
价格变动历史
- 输入 $0.10 → $0.20 +100% · 输出 $0.34 → $0.40 +18%
- 输入 $0.20 → $0.10 -50% · 输出 $0.40 → $0.34 -15%
OpenRouter 上的服务商
当前有 16 个接口提供此模型。正常运行率为 OpenRouter 上的请求成功率；价格单位为每 1M token。
| 服务商 | 输入 | 输出 | 上下文窗口 | 30 分钟正常率 | 24 小时正常率 |
|---|---|---|---|---|---|
| DeepInfrafp8 · flex优惠/优先级通道 | $0.16 | $0.32 | 262K | 100% | 96.3% |
| CoreWeavefp4 | $0.10 | $0.34 | 262K | 100% | 99.72% |
| Venicefp4 | $0.12 | $0.36 | 256K | 100% | 100% |
| Chutesfp4 | $0.12 | $0.37 | 131K | 90.0% | 99.65% |
| Friendli | $0.14 | $0.40 | 262K | 100% | 100% |
| DekaLLM | $0.14 | $0.40 | 262K | 100% | 99.63% |
| Crusoebf16 | $0.14 | $0.40 | 262K | 100% | 99.38% |
| Novitabf16 | $0.14 | $0.40 | 262K | 100% | 87.4% |
| Parasailfp8 | $0.15 | $0.40 | 262K | 100% | 100% |
| DeepInfrafp8 | $0.20 | $0.40 | 262K | 100% | 100% |
| DeepInfraturbo | $0.135 | $0.51 | 262K | 100% | 100% |
| DeepInfrafp8 · priority优惠/优先级通道 | $0.30 | $0.60 | 262K | 100% | 99.93% |
| ModelRunfp4 | $0.75 | $1.00 | 262K | 100% | 100% |
| SiliconFlowfp8 | $0.75 | $1.00 | 262K | 100% | 99.41% |
| Io Net | $0.361 | $1.09 | 262K | 100% | 98.8% |
| SambaNova | $0.38 | $1.15 | 131K | 100% | 93.7% |