OpenCode 榜单 — AI 编程模型用量与成本排名
OpenCode 榜单
用量时序、成本、缓存、份额、地区、留存 · 对齐源站全部板块
用量第一
DeepSeek V4.1 Flash
4.0T tokens
最便宜
Space Bunny
$0.00/次
涨最快
—
份额第一
DeepSeek
58.3%
Top Models · 时序
Leaderboard · 模型用量榜
· 点击表头排序#
模型
会话成本为 opencode 实测的单次任务中位数开销，与用量榜相互独立——部分模型源站未提供成本数据（显示 —），并非缺失或故障。
Session Cost · 会话成本
· 单次任务中位数成本，最省在前| # | 模型 | $/会话 | Tokens |
|---|---|---|---|
| 1 | Space Bunny | $0.00 | 587.2T |
| 2 | MIMO V2.6 Flash | $0.00 | 1238.4T |
| 3 | Muse Spark 1.3 Contributor | $0.00 | 2074.9T |
| 4 | DeepSeek V4 Flash | $0.03 | 3362.1T |
| 5 | DeepSeek V4.1 Flash | $0.05 | 3811.7T |
| 6 | GLM 5.3 Flash | $0.08 | 2030.6T |
Token Cost · Token 价格
· 每百万 token 单价（$/M），最省在前| # | 模型 | 加权 | 输入 | 输出 | 缓存 |
|---|---|---|---|---|---|
| 1 | Muse Spark 1.3 Contributor | $0.200 | $0.100 | $0.200 | $0.002 |
| 2 | MIMO V2.6 Flash | $0.320 | $0.160 | $0.320 | $0.004 |
| 3 | GLM 5.3 Flash | $0.500 | $0.150 | $0.500 | $0.030 |
| 4 | DeepSeek V4 Flash | $1.200 | $0.300 | $1.200 | $0.006 |
| 5 | DeepSeek V4.1 Flash | $1.200 | $0.300 | $1.200 | $0.006 |
| 6 | Longcat 2.5 Preview | $1.200 | $0.300 | $1.200 | $0.006 |
| 7 | Step 5 Preview | $2.741 | $0.959 | $2.741 | $0.048 |
Cache Ratio · 缓存命中率
· 命中越高成本越省| # | 模型 | 缓存率 | 命中 / 未命中 |
|---|---|---|---|
| 1 | Space Bunny | 96.5% | 20.8k / 757.9 |
| 2 | DeepSeek V4.1 Flash | 96.2% | 42.4k / 1.7k |
| 3 | Muse Spark 1.3 Contributor | 95.7% | 34.8k / 1.6k |
| 4 | Step 5 Preview | 95.5% | 4.8k / 223.8 |
| 5 | DeepSeek V4 Flash | 94.6% | 4.6k / 262.7 |
| 6 | MIMO V2.6 Flash | 94.1% | 10.8k / 678.7 |
| 7 | DeepSeek V4 1 Flash Opencode Bo44an | 93.7% | 849 / 57.3 |
| 8 | GLM 5.3 Flash | 92.1% | 2.0k / 170.3 |
| 9 | Longcat 2.5 Preview | 92.1% | 1.8k / 155.5 |
| 10 | Nemotron 3 Ultra | 90.3% | 3.8k / 401.8 |
Retention · 用户留存榜
· 模型周留存率（回访 / 可回访）| # | 模型 | 留存率 | 样本周 |
|---|---|---|---|
| 1 | MIMO V2.6 Flash Xiaomi | 80.3% | 2.8k |
| 2 | Space Bunny Unknown | 80% | 8.2k |
| 3 | Muse Spark 1.2 Contributor Meta | 78.3% | 9.1k |
| 4 | Kimi K2.5 Moonshot | 77.1% | 214 |
| 5 | MIMO V2.5 Xiaomi | 72.1% | 26.4k |
| 6 | Qwen3.5 Plus Qwen | 71.9% | 139 |
| 7 | Hy3 Tencent | 71.5% | 3.4k |
| 8 | Minimax M2.5 MiniMax | 69.4% | 245 |
| 9 | DeepSeek V4 Flash DeepSeek | 69.3% | 272.5k |
| 10 | Qwen3.7 Plus Qwen | 67.1% | 8.4k |
| 11 | Muse Spark 1.3 Contributor Meta | 67% | 41.2k |
| 12 | GLM 5 Zhipu | 66.7% | 384 |
| 13 | Minimax M2.7 MiniMax | 66.1% | 2.2k |
| 14 | MIMO V2.5 Pro Xiaomi | 65.7% | 5.1k |
| 15 | GPT 6 Luna OpenAI | 65.2% | 3.5k |