价格变动时间线
各家大模型 API 的调价 / 降价记录,按日期倒序 · 数据更新于 2026-11-15
📊 全部数据 · 总览每日自动比对各模型价格,有变动即记录在此。绿色为降价、红色为涨价; 点击模型名查看该模型完整价格历史曲线。← 返回总表
2026-11-15
- DeepSeek: R1 0528最大输出32K → 33K
- DeepSeek: R1 0528输出价$2.29 → $2.15-6%
- DeepSeek: R1 0528输入价$0.57 → $0.50-12%
- OpenAI: gpt-oss-20b最大输出118K → 66K
- OpenAI: gpt-oss-20b缓存读$0.03 → $0.04+25%
- OpenAI: gpt-oss-20b输出价$0.13 → $0.30+131%
- OpenAI: gpt-oss-20b输入价$0.03 → $0.07+150%
- Z.ai: GLM 4.7缓存读$0.11 → $0.08-27%
- Z.ai: GLM 4.7输出价$2.20 → $1.75-20%
- Z.ai: GLM 4.7输入价$0.60 → $0.40-33%
- Qwen: Qwen3.5-9B缓存读$0.08 → $0.12+38%
- Qwen: Qwen3.5-9B输出价$0.14 → $0.19+39%
- Qwen: Qwen3.5-9B输入价$0.09 → $0.13+39%
- OpenAI: GPT-5.6 Sol缓存写$5.00 → $5.50+10%
- OpenAI: GPT-5.6 Sol缓存读$0.40 → $0.44+10%
- OpenAI: GPT-5.6 Sol输出价$20.00 → $22.00+10%
- OpenAI: GPT-5.6 Sol输入价$4.00 → $4.40+10%
- IBM: Granite 4.2 8B输出价$0.25 → $0.15-40%
- DeepSeek: DeepSeek V4.1 Flash缓存读$0.01 → $0.00-83%
- DeepSeek: DeepSeek V4.1 Flash输出价$1.20 → $0.20-83%
- DeepSeek: DeepSeek V4.1 Flash输入价$0.30 → $0.05-83%
- NVIDIA: Nemotron 3.5 Super VL最大输出944K → 236K
- NVIDIA: Nemotron 3.5 Super VL输出价$0.47 → $0.50+6%
- DeepSeek: DeepSeek V4.1 Flash 1017最大输出944K → 384K
- DeepSeek: DeepSeek V4.1 Flash 1017缓存读$0.01 → $0.01-30%
- DeepSeek: DeepSeek V4.1 Flash 1017输出价$0.50 → $0.11-78%
- DeepSeek: DeepSeek V4.1 Flash 1017输入价$0.04 → $0.02-47%
2026-11-14
- DeepSeek: R1 0528最大输出33K → 32K
- DeepSeek: R1 0528输出价$2.15 → $2.29+6%
- DeepSeek: R1 0528输入价$0.50 → $0.57+14%
- Qwen: Qwen3.5-9B缓存读$0.13 → $0.08-37%
- Qwen: Qwen3.5-9B输出价$0.22 → $0.14-37%
- Qwen: Qwen3.5-9B输入价$0.15 → $0.09-38%
- Z.ai: GLM 5.1最大输出131K → 182K
- Xiaomi: MiMo-V2.5-Pro缓存读$0.10 → $0.00-96%
- OpenAI: GPT-5.6 Sol缓存写$5.50 → $5.00-9%
- OpenAI: GPT-5.6 Sol缓存读$0.44 → $0.40-9%
- OpenAI: GPT-5.6 Sol输出价$22.00 → $20.00-9%
- OpenAI: GPT-5.6 Sol输入价$4.40 → $4.00-9%
- Qwen: Qwen3.8 27B上下文窗口1.05M → 1.02M
- DeepSeek: DeepSeek V4.1 Flash最大输出944K → 393K
- DeepSeek: DeepSeek V4.1 Flash 1017最大输出393K → 944K
- DeepSeek: DeepSeek V4.1 Flash 1017缓存读$0.00 → $0.01+233%
- DeepSeek: DeepSeek V4.1 Flash 1017输出价$0.60 → $0.50-17%
- DeepSeek: DeepSeek V4.1 Flash 1017输入价$0.15 → $0.04-75%
- Particle AI: Particle-1输入价$0.05 → $0.06+20%
2026-11-13
- Meta: Llama 3.2 3B Instruct最大输出118K → 72K
- Meta: Llama 3.2 3B Instruct输出价$0.33 → $0.34+2%
- Meta: Llama 3.2 3B Instruct输入价$0.05 → $0.05+2%
- DeepSeek: DeepSeek V3最大输出16K → 16K
- DeepSeek: DeepSeek V3输出价$0.89 → $1.03+16%
- DeepSeek: DeepSeek V3输入价$0.32 → $0.26-20%
- OpenAI: gpt-oss-20b缓存读$0.01 → $0.03+233%
- OpenAI: gpt-oss-20b输出价$0.09 → $0.13+44%
- OpenAI: gpt-oss-20b输入价$0.02 → $0.03+67%
- MiniMax: MiniMax M2.5最大输出128K → 33K
- MiniMax: MiniMax M2.5上下文窗口200K → 198K
- MiniMax: MiniMax M2.5缓存读$0.03 → $0.03+11%
- MiniMax: MiniMax M2.5输出价$1.08 → $0.95-12%
- Qwen: Qwen3.5-9B缓存读$0.14 → $0.13-8%
- Qwen: Qwen3.5-9B输出价$0.24 → $0.22-8%
- Qwen: Qwen3.5-9B输入价$0.16 → $0.15-7%
- Google: Gemma 4 26B A4B 缓存读$0.04 → $0.04-12%
- Google: Gemma 4 26B A4B 输出价$0.26 → $0.23-12%
- Google: Gemma 4 26B A4B 输入价$0.08 → $0.07-12%
- Xiaomi: MiMo-V2.5-Pro最大输出236K → 131K
- Xiaomi: MiMo-V2.5-Pro输出价$1.80 → $0.87-52%
- Xiaomi: MiMo-V2.5-Pro输入价$0.48 → $0.43-9%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)最大输出384K → 393K
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.07 → $0.25+237%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.78 → $1.50-16%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.89 → $0.90+1%
- Z.ai: GLM 5.3缓存读$0.07 → $0.08+15%
- Z.ai: GLM 5.3输出价$1.40 → $3.60+157%
- Z.ai: GLM 5.3输入价$0.07 → $0.08+20%
- IBM: Granite 4.2 8B输出价$0.15 → $0.25+67%
- DeepSeek: DeepSeek V4.1 Flash输入价$0.05 → $0.30+488%
- Meta: Muse Spark Open最大输出262K → 944K
- DeepSeek: DeepSeek V4.1 Flash 1017输入价$0.01 → $0.15+2400%
- MoonshotAI: Kimi K3.1 Pro输出价$8.41 → $10.00+19%
- Ricursive: Ripple 65B Preview缓存读$0.05 → $0.03-50%
- Ricursive: Ripple 65B Preview输出价$0.60 → $0.30-50%
- Ricursive: Ripple 65B Preview输入价$0.20 → $0.10-50%
2026-11-12
- bytedance-seed/seed-1.6输出价$2.00 → $0.00-100%
- bytedance-seed/seed-1.6输入价$0.25 → $0.00-100%
- bytedance-seed/seed-1.6-flash输出价$0.30 → $0.00-100%
- bytedance-seed/seed-1.6-flash输入价$0.07 → $0.00-100%
- Qwen: Qwen3 14B最大输出37K → 16K
- Qwen: Qwen3 14B输出价$0.22 → $0.24+9%
- Qwen: Qwen3 14B输入价$0.10 → $0.12+20%
- OpenAI: gpt-oss-20b缓存读$0.03 → $0.01-70%
- OpenAI: gpt-oss-20b输出价$0.13 → $0.09-31%
- OpenAI: gpt-oss-20b输入价$0.03 → $0.02-40%
- MoonshotAI: Kimi K2.5最大输出66K → 236K
- MoonshotAI: Kimi K2.5缓存读$0.21 → $0.15-28%
- MoonshotAI: Kimi K2.5输出价$3.33 → $2.50-25%
- MoonshotAI: Kimi K2.5输入价$0.53 → $0.50-6%
- Qwen: Qwen3.5-9B最大输出236K → 41K
- Qwen: Qwen3.5-9B缓存读— → $0.14
- Qwen: Qwen3.5-9B输出价$0.25 → $0.24-5%
- Qwen: Qwen3.5-9B输入价$0.17 → $0.16-5%
- Google: Gemma 4 26B A4B 最大输出128K → 236K
- Google: Gemma 4 26B A4B 缓存读$0.05 → $0.04-8%
- Google: Gemma 4 26B A4B 输出价$0.10 → $0.26+155%
- Google: Gemma 4 26B A4B 输入价$0.08 → $0.08-4%
- Z.ai: GLM 5.1缓存读$0.25 → $0.26+5%
- Z.ai: GLM 5.1输出价$4.18 → $4.40+5%
- Z.ai: GLM 5.1输入价$1.33 → $1.40+5%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.07 → $0.07+2%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.75 → $1.78+2%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.88 → $0.89+2%
- OpenAI: GPT-5.6 Sol缓存写$5.00 → $5.50+10%
- OpenAI: GPT-5.6 Sol缓存读$0.40 → $0.44+10%
- OpenAI: GPT-5.6 Sol输出价$20.00 → $22.00+10%
- OpenAI: GPT-5.6 Sol输入价$4.00 → $4.40+10%
- IBM: Granite 4.2 8B输出价$0.25 → $0.15-40%
- DeepSeek: DeepSeek V4.1 Flash缓存读$0.02 → $0.01-67%
- DeepSeek: DeepSeek V4.1 Flash输出价$0.57 → $1.20+112%
- DeepSeek: DeepSeek V4.1 Flash 1017最大输出944K → 393K
- DeepSeek: DeepSeek V4.1 Flash 1017缓存读$0.01 → $0.00-40%
- DeepSeek: DeepSeek V4.1 Flash 1017输出价$0.50 → $0.60+20%
- MoonshotAI: Kimi K3.1 Pro输出价$10.00 → $8.41-16%
2026-11-11
- bytedance-seed/seed-1.6最大输出33K → —
- bytedance-seed/seed-1.6上下文窗口262K → —
- bytedance-seed/seed-1.6-flash最大输出33K → —
- bytedance-seed/seed-1.6-flash上下文窗口262K → —
- Google: Gemma 3 27B输出价$0.45 → $0.35-22%
- MiniMax: MiniMax M2.5最大输出59K → 128K
- MiniMax: MiniMax M2.5缓存读$0.06 → $0.03-55%
- MiniMax: MiniMax M2.5输出价$1.20 → $1.08-10%
- MiniMax: MiniMax M2.5输入价$0.30 → $0.27-10%
- NVIDIA: Nemotron 3 Super上下文窗口1.05M → 262K
- Z.ai: GLM 5.1缓存读$0.26 → $0.25-5%
- Z.ai: GLM 5.1输出价$4.40 → $4.18-5%
- Z.ai: GLM 5.1输入价$1.40 → $1.33-5%
- DeepSeek: DeepSeek V4 Flash 0423最大输出944K → 393K
- DeepSeek: DeepSeek V4 Flash 0423缓存读$0.02 → $0.02-1%
- DeepSeek: DeepSeek V4 Flash 0423输出价$0.18 → $0.18-1%
- DeepSeek: DeepSeek V4 Flash 0423输入价$0.09 → $0.09-1%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.07 → $0.07-1%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.77 → $1.75-1%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.88 → $0.88-1%
- Z.ai: GLM 5.3 Flash输出价$0.50 → $0.08-84%
- DeepSeek: DeepSeek V4.1 Flash缓存读$0.01 → $0.02+200%
- DeepSeek: DeepSeek V4.1 Flash输出价$1.20 → $0.57-53%
2026-11-10
- ReMM SLERP 13B缓存读$0.35 → $0.20-43%
- ReMM SLERP 13B输出价$0.65 → $0.50-23%
- ReMM SLERP 13B输入价$0.35 → $0.30-14%
- Mancer: Weaver (alpha)缓存读$0.40 → $0.30-25%
- Mancer: Weaver (alpha)输出价$0.75 → $0.50-33%
- Mancer: Weaver (alpha)输入价$0.40 → $0.30-25%
- Meta: Llama 3.3 70B Instruct最大输出33K → 115K
- Meta: Llama 3.3 70B Instruct缓存读$0.29 → $0.71+141%
- Meta: Llama 3.3 70B Instruct输出价$0.79 → $0.71-10%
- Meta: Llama 3.3 70B Instruct输入价$0.59 → $0.71+20%
- Qwen: Qwen3 14B最大输出16K → 37K
- Qwen: Qwen3 14B输出价$0.24 → $0.22-8%
- Qwen: Qwen3 14B输入价$0.12 → $0.10-17%
- Google: Gemma 4 26B A4B 最大输出236K → 128K
- Google: Gemma 4 26B A4B 缓存读$0.04 → $0.05+8%
- Google: Gemma 4 26B A4B 输出价$0.26 → $0.10-61%
- Google: Gemma 4 26B A4B 输入价$0.08 → $0.08+5%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.07 → $0.07+1%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.75 → $1.77+1%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.88 → $0.88+1%
- Z.ai: GLM 5.2缓存读$0.01 → $0.01+20%
- Z.ai: GLM 5.2输入价$0.01 → $0.01+17%
- Qwen: Qwen3.8 27B上下文窗口1.02M → 1.05M
- Z.ai: GLM 5.3 Flash最大输出944K → 131K
- Z.ai: GLM 5.3 Flash输入价$0.15 → $0.04-73%
- DeepSeek: DeepSeek V4.1 Flash缓存读$0.02 → $0.01-67%
- DeepSeek: DeepSeek V4.1 Flash输出价$0.57 → $1.20+112%
- Meta: Muse Spark Open最大输出131K → 262K
2026-11-09
- Auto Router (Beta)最大输出66K → 64K
- TypeSafe: Jev Router最大输出66K → 64K
- Auto Code最大输出66K → 64K
- Meta: Llama 3.1 8B Instruct缓存读$0.03 → —
- Google: Gemma 3 27B输出价$0.35 → $0.45+29%
- Mistral: Mistral Small 3.2 24B最大输出16K → 26K
- Mistral: Mistral Small 3.2 24B缓存读— → $0.01
- Mistral: Mistral Small 3.2 24B输出价$0.20 → $0.30+50%
- Mistral: Mistral Small 3.2 24B输入价$0.07 → $0.10+33%
- Z.ai: GLM 4.5 Air缓存读$0.03 → $0.03+20%
- Z.ai: GLM 4.5 Air输出价$0.85 → $1.10+29%
- Z.ai: GLM 4.5 Air输入价$0.13 → $0.20+54%
- MiniMax: MiniMax M2.5上下文窗口205K → 200K
- Qwen: Qwen3.5 397B A17B最大输出236K → 64K
- Qwen: Qwen3.5 397B A17B缓存读$0.30 → $0.12-60%
- Qwen: Qwen3.5 397B A17B输入价$0.50 → $0.60+20%
- MiniMax: MiniMax M2.7上下文窗口205K → 197K
- Google: Gemma 4 26B A4B 缓存读$0.04 → $0.04+13%
- Google: Gemma 4 26B A4B 输出价$0.23 → $0.26+13%
- Google: Gemma 4 26B A4B 输入价$0.07 → $0.08+13%
- Xiaomi: MiMo-V2.5-Pro上下文窗口1.05M → 1.02M
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.02 → $0.07+319%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$0.42 → $1.75+319%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.21 → $0.88+319%
- Z.ai: GLM 5.3 Flash最大输出131K → 944K
- Z.ai: GLM 5.3 Flash输入价$0.03 → $0.15+348%
- StepFun: Step 5 Preview上下文窗口1.05M → 1M
- NVIDIA: Nemotron 3.5 Super VL输出价$0.50 → $0.47-6%
- Meta: Muse Spark Open最大输出944K → 131K
- DeepSeek: DeepSeek V4.1 Flash 1017输入价$0.15 → $0.01-96%
2026-11-08
- Meta: Llama 3.1 8B Instruct最大输出118K → 16K
- Meta: Llama 3.1 8B Instruct输出价$0.08 → $0.04-50%
- Meta: Llama 3.1 8B Instruct输入价$0.05 → $0.02-60%
- Meta: Llama 3.2 3B Instruct最大输出72K → 118K
- Meta: Llama 3.2 3B Instruct输出价$0.34 → $0.33-1%
- Meta: Llama 3.2 3B Instruct输入价$0.05 → $0.05-2%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.04 → $0.02-59%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.01 → $0.42-59%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.51 → $0.21-59%
- Qwen: Qwen3.6 27B最大输出66K → 82K
- Qwen: Qwen3.6 27B输出价$2.70 → $3.20+19%
- Qwen: Qwen3.6 27B输入价$0.45 → $0.32-29%
- DeepSeek: DeepSeek V4 Pro 0813缓存读$0.11 → $0.10-8%
- Z.ai: GLM 5.3 Flash最大输出944K → 131K
- Z.ai: GLM 5.3 Flash输入价$0.15 → $0.03-78%
- DeepSeek: DeepSeek V4.1 Flash 1017最大输出393K → 944K
- DeepSeek: DeepSeek V4.1 Flash 1017缓存读$0.00 → $0.01+67%
- DeepSeek: DeepSeek V4.1 Flash 1017输出价$0.60 → $0.50-17%
2026-11-07
- ReMM SLERP 13B缓存读— → $0.35
- Mancer: Weaver (alpha)缓存读— → $0.40
- Magnum v4 72B缓存读— → $2.50
- MoonshotAI: Kimi K2.5最大输出236K → 66K
- MoonshotAI: Kimi K2.5缓存读$0.15 → $0.21+39%
- MoonshotAI: Kimi K2.5输出价$2.50 → $3.33+33%
- MoonshotAI: Kimi K2.5输入价$0.50 → $0.53+6%
- MiniMax: MiniMax M2.5最大输出33K → 59K
- MiniMax: MiniMax M2.5缓存读$0.03 → $0.06+100%
- MiniMax: MiniMax M2.5输出价$0.95 → $1.20+26%
- MiniMax: MiniMax M2.5输入价$0.27 → $0.30+11%
- NVIDIA: Nemotron 3 Super缓存读$0.15 → —
- NVIDIA: Nemotron 3 Super输出价$2.40 → $0.40-83%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.06 → $0.04-29%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.42 → $1.01-29%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.71 → $0.51-29%
- DeepSeek: DeepSeek V4 Pro 0813缓存读$0.11 → $0.11-5%
- DeepSeek: DeepSeek V4 Pro 0813输出价$4.66 → $4.44-5%
- DeepSeek: DeepSeek V4 Pro 0813输入价$0.12 → $0.11-5%
- inclusionAI: Ling 3.1 Flash缓存读$0.01 → $0.01-40%
- inclusionAI: Ling 3.1 Flash输出价$0.23 → $0.14-40%
- inclusionAI: Ling 3.1 Flash输入价$0.07 → $0.04-40%
2026-11-06
- Meta: Llama 3.3 70B Instruct缓存读$0.10 → $0.29+195%
- Meta: Llama 3.3 70B Instruct输入价$0.10 → $0.59+490%
- DeepSeek: DeepSeek V3最大输出16K → 16K
- DeepSeek: DeepSeek V3输出价$1.03 → $0.89-13%
- DeepSeek: DeepSeek V3输入价$0.26 → $0.32+24%
- Qwen: Qwen3 Next 80B A3B Instruct缓存读$0.07 → —
- NVIDIA: Nemotron 3 Super最大输出262K → 16K
- NVIDIA: Nemotron 3 Super输入价$0.30 → $0.07-75%
- Google: Gemma 4 26B A4B 缓存读$0.04 → $0.04-12%
- Google: Gemma 4 26B A4B 输出价$0.26 → $0.23-12%
- Google: Gemma 4 26B A4B 输入价$0.08 → $0.07-12%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)最大输出393K → 384K
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.14 → $0.06-56%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$3.20 → $1.42-56%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$1.60 → $0.71-56%
- DeepSeek: DeepSeek V4 Pro 0813缓存读$0.10 → $0.11+13%
- DeepSeek: DeepSeek V4 Pro 0813输出价$4.11 → $4.66+13%
- DeepSeek: DeepSeek V4 Pro 0813输入价$0.10 → $0.12+13%
- DeepSeek: DeepSeek V4 Flash Vision Exp最大输出262K → 944K
- DeepSeek: DeepSeek V4 Flash Vision Exp缓存读$0.01 → $0.01+103%
- DeepSeek: DeepSeek V4 Flash Vision Exp输出价$0.65 → $1.32+104%
- DeepSeek: DeepSeek V4 Flash Vision Exp输入价$0.22 → $0.44+104%
- DeepSeek: DeepSeek V4.1 Flash输出价$0.50 → $0.57+13%
- DeepSeek: DeepSeek V4.1 Flash输入价$0.04 → $0.05+13%
2026-11-05
- Meta: Llama 3.3 70B Instruct最大输出16K → 33K
- Meta: Llama 3.3 70B Instruct输出价$0.32 → $0.79+147%
- Meta: Llama 4 Maverick最大输出16K → 33K
- Meta: Llama 4 Maverick缓存读$0.05 → $0.17+240%
- Meta: Llama 4 Maverick输出价$0.65 → $1.00+53%
- Meta: Llama 4 Maverick输入价$0.19 → $0.35+87%
- OpenAI: gpt-oss-20b最大输出33K → 118K
- OpenAI: gpt-oss-20b缓存读$0.01 → $0.03+233%
- OpenAI: gpt-oss-20b输出价$0.09 → $0.13+44%
- OpenAI: gpt-oss-20b输入价$0.02 → $0.03+67%
- OpenAI: gpt-oss-120b缓存读$0.02 → $0.07+275%
- Qwen: Qwen3 Next 80B A3B Instruct最大输出236K → 16K
- Qwen: Qwen3 Next 80B A3B Instruct输入价$0.10 → $0.09-10%
- MiniMax: MiniMax M2.7最大输出131K → 177K
- MiniMax: MiniMax M2.7缓存读$0.06 → $0.04-30%
- MiniMax: MiniMax M2.7输出价$1.20 → $0.84-30%
- MiniMax: MiniMax M2.7输入价$0.30 → $0.21-30%
- Google: Gemma 4 26B A4B 最大输出33K → 236K
- Google: Gemma 4 26B A4B 缓存读$0.02 → $0.04+166%
- Google: Gemma 4 26B A4B 输出价$0.10 → $0.26+160%
- Google: Gemma 4 26B A4B 输入价$0.03 → $0.08+132%
- MoonshotAI: Kimi K2.7 Code缓存读$0.15 → $0.16+7%
- DeepSeek: DeepSeek V4 Pro 0813缓存读$0.11 → $0.10-11%
- DeepSeek: DeepSeek V4 Pro 0813输出价$4.60 → $4.11-11%
- DeepSeek: DeepSeek V4 Pro 0813输入价$0.12 → $0.10-10%
- Z.ai: GLM 5.3 Flash输出价$0.08 → $0.50+491%
- Inference.net: Schematron V2 Small最大输出4K → 8K
- Reflection: Beam 1.0 501B A23B最大输出131K → 944K
- Reflection: Beam 1.0 501B A23B缓存读$0.05 → $0.04-20%
- Reflection: Beam 1.0 501B A23B输入价$0.25 → $0.30+20%
- ricursiveAI/Ripple-65B-Preview最大输出15K → 131K
- ricursiveAI/Ripple-65B-Preview上下文窗口16K → 1.05M
2026-11-04
- Meta: Llama 3.3 70B Instruct最大输出128K → 16K
- Meta: Llama 3.3 70B Instruct输出价$0.52 → $0.32-38%
- Meta: Llama 3.3 70B Instruct输入价$0.20 → $0.10-50%
- OpenAI: gpt-oss-20b最大输出118K → 33K
- OpenAI: gpt-oss-20b缓存读$0.03 → $0.01-70%
- OpenAI: gpt-oss-20b输出价$0.13 → $0.09-31%
- OpenAI: gpt-oss-20b输入价$0.03 → $0.02-40%
- OpenAI: gpt-oss-120b缓存读$0.02 → $0.02-9%
- Google: Gemma 4 26B A4B 缓存读$0.03 → $0.02-52%
- Google: Gemma 4 26B A4B 输出价$0.20 → $0.10-51%
- Google: Gemma 4 26B A4B 输入价$0.07 → $0.03-50%
- Xiaomi: MiMo-V2.5上下文窗口1.05M → 1M
- DeepSeek: DeepSeek V4 Flash 0731输入价$0.06 → $0.04-26%
- DeepSeek: DeepSeek V4 Pro 0813输出价$4.00 → $4.60+15%
- Z.ai: GLM 5.3缓存读$0.08 → $0.07-9%
- Z.ai: GLM 5.3输出价$3.60 → $1.40-61%
- Z.ai: GLM 5.3输入价$0.08 → $0.07-13%
- Z.ai: GLM 5.3 Flash最大输出131K → 944K
- Z.ai: GLM 5.3 Flash输入价$0.03 → $0.15+379%
- IBM: Granite 4.2 8B输出价$0.15 → $0.25+67%
- DeepSeek: DeepSeek V4.1 Flash缓存读$0.01 → $0.02+50%
2026-11-03
- Meta: Llama 3.1 70B Instruct最大输出16K → 8K
- Meta: Llama 3.1 70B Instruct输出价$0.40 → $0.72+80%
- Meta: Llama 3.1 70B Instruct输入价$0.40 → $0.72+80%
- Meta: Llama 3.3 70B Instruct最大输出16K → 128K
- Meta: Llama 3.3 70B Instruct缓存读— → $0.10
- Meta: Llama 3.3 70B Instruct输出价$0.32 → $0.52+63%
- Meta: Llama 3.3 70B Instruct输入价$0.10 → $0.20+100%
- Qwen: Qwen3 235B A22B Instruct 2507最大输出16K → 118K
- Qwen: Qwen3 235B A22B Instruct 2507缓存读— → $0.05
- Qwen: Qwen3 235B A22B Instruct 2507输出价$0.55 → $0.80+45%
- Qwen: Qwen3 235B A22B Instruct 2507输入价$0.09 → $0.14+56%
- Qwen: Qwen3 Coder 30B A3B Instruct输出价$0.28 → $0.60+114%
- Qwen: Qwen3 Coder 30B A3B Instruct输入价$0.07 → $0.15+114%
- OpenAI: gpt-oss-120b缓存读$0.07 → $0.02-71%
- Qwen: Qwen3 Next 80B A3B Instruct最大输出16K → 236K
- Qwen: Qwen3 Next 80B A3B Instruct缓存读— → $0.07
- Qwen: Qwen3 Next 80B A3B Instruct输入价$0.09 → $0.10+11%
- Google: Gemma 4 26B A4B 最大输出236K → 33K
- Google: Gemma 4 26B A4B 缓存读$0.04 → $0.03-22%
- Google: Gemma 4 26B A4B 输出价$0.26 → $0.20-22%
- Google: Gemma 4 26B A4B 输入价$0.08 → $0.07-14%
- MoonshotAI: Kimi K3输出价$15.00 → $16.00+7%
- DeepSeek: DeepSeek V4 Flash 0731缓存读$0.02 → $0.04+47%
- DeepSeek: DeepSeek V4 Flash 0731输出价$0.24 → $0.30+26%
- DeepSeek: DeepSeek V4 Flash 0731输入价$0.12 → $0.06-50%
- Z.ai: GLM 5.3 Flash缓存读$0.03 → $0.03+5%
- Z.ai: GLM 5.3 Flash输出价$0.09 → $0.08-8%
- Z.ai: GLM 5.3 Flash输入价$0.03 → $0.03+10%
- DeepSeek: DeepSeek V4.1 Flash缓存读$0.01 → $0.01+100%
- NVIDIA: Nemotron 3.5 Super VL最大输出236K → 944K
2026-11-02
- Meta: Llama 3.2 3B Instruct最大输出118K → 72K
- Meta: Llama 3.2 3B Instruct输出价$0.33 → $0.34+2%
- Meta: Llama 3.2 3B Instruct输入价$0.05 → $0.05+2%
- Google: Gemma 4 26B A4B 缓存读$0.04 → $0.04+13%
- Google: Gemma 4 26B A4B 输出价$0.23 → $0.26+13%
- Google: Gemma 4 26B A4B 输入价$0.07 → $0.08+13%
- Qwen: Qwen3.8 27B上下文窗口1M → 1.02M
- Z.ai: GLM 5.3 Flash输入价$0.15 → $0.03-81%
- Z.ai: GLM 5.3 FlashX缓存读$0.09 → $0.07-17%
- Meta: Muse Spark Open最大输出131K → 944K
2026-11-01
- OpenAI: gpt-oss-20b最大输出33K → 118K
- OpenAI: gpt-oss-20b缓存读$0.01 → $0.03+233%
- OpenAI: gpt-oss-20b输出价$0.09 → $0.13+44%
- OpenAI: gpt-oss-20b输入价$0.02 → $0.03+67%
- DeepSeek: DeepSeek V4 Pro 0813缓存读$0.02 → $0.11+400%
- DeepSeek: DeepSeek V4 Pro 0813输入价$0.66 → $0.12-83%
- Qwen: Qwen3.8 27B缓存读$0.02 → $0.04+87%
- Qwen: Qwen3.8 27B输出价$2.30 → $1.88-18%
- Qwen: Qwen3.8 27B输入价$0.04 → $0.15+275%
- Z.ai: GLM 5.3缓存读$0.07 → $0.08+10%
- Z.ai: GLM 5.3输出价$1.40 → $3.60+157%
- Z.ai: GLM 5.3输入价$0.07 → $0.08+14%
- Z.ai: GLM 5.3 Flash缓存读$0.01 → $0.03+91%
- Z.ai: GLM 5.3 Flash输出价$0.15 → $0.09-40%
2026-10-31
- Z.ai: GLM 4.7缓存读$0.10 → $0.11+11%
- Z.ai: GLM 4.7输出价$1.98 → $2.20+11%
- Z.ai: GLM 4.7输入价$0.54 → $0.60+11%
- MiniMax: MiniMax M2.5最大输出131K → 33K
- MiniMax: MiniMax M2.5输出价$1.20 → $0.95-21%
- MiniMax: MiniMax M2.5输入价$0.30 → $0.27-10%
- Qwen: Qwen3.6 27B最大输出82K → 66K
- Qwen: Qwen3.6 27B输出价$3.20 → $2.70-16%
- Qwen: Qwen3.6 27B输入价$0.32 → $0.45+41%
- Z.ai: GLM 5.2输出价$7.00 → $6.80-3%
- DeepSeek: DeepSeek V4 Pro 0813最大输出393K → 944K
- DeepSeek: DeepSeek V4 Pro 0813输出价$1.98 → $4.00+102%
- Qwen: Qwen3.8 27B缓存读$0.04 → $0.02-47%
- Qwen: Qwen3.8 27B输出价$1.88 → $2.30+23%
- Qwen: Qwen3.8 27B输入价$0.15 → $0.04-73%
- Z.ai: GLM 5.3缓存读$0.10 → $0.07-30%
- Z.ai: GLM 5.3输出价$2.77 → $1.40-49%
- Z.ai: GLM 5.3输入价$0.10 → $0.07-33%
- Z.ai: GLM 5.3 Flash缓存读$0.02 → $0.01-24%
- Z.ai: GLM 5.3 Flash输出价$0.13 → $0.15+17%
- Particle AI: Particle-1缓存读$0.03 → $0.00-90%
- Particle AI: Particle-1输入价$0.25 → $0.05-80%
2026-10-30
- DeepSeek: DeepSeek V3.2最大输出66K → 147K
- DeepSeek: DeepSeek V3.2缓存读$0.03 → $0.13+364%
- DeepSeek: DeepSeek V3.2输出价$0.42 → $0.38-10%
- DeepSeek: DeepSeek V3.2输入价$0.28 → $0.26-7%
- Z.ai: GLM 4.6V缓存读$0.05 → $0.06+10%
- Google: Gemma 4 26B A4B 缓存读$0.04 → $0.04-12%
- Google: Gemma 4 26B A4B 输出价$0.26 → $0.23-12%
- Google: Gemma 4 26B A4B 输入价$0.08 → $0.07-12%
- Qwen: Qwen3.6 27B最大输出66K → 82K
- Qwen: Qwen3.6 27B输出价$2.70 → $3.20+19%
- Qwen: Qwen3.6 27B输入价$0.45 → $0.32-29%
- MoonshotAI: Kimi K2.7 Code输出价$3.10 → $3.50+13%
- MoonshotAI: Kimi K2.7 Code输入价$0.56 → $0.60+7%
- MoonshotAI: Kimi K3缓存读$0.21 → $0.18-13%
- MoonshotAI: Kimi K3输出价$10.80 → $15.00+39%
- MoonshotAI: Kimi K3输入价$0.24 → $0.19-20%
- Qwen: Qwen3.8 2.4T A95B最大输出909K → 131K
- Qwen: Qwen3.8 27B最大输出262K → 236K
- Qwen: Qwen3.8 27B缓存读$0.05 → $0.04-25%
- Qwen: Qwen3.8 27B输出价$2.50 → $1.88-25%
- Qwen: Qwen3.8 27B输入价$0.20 → $0.15-25%
- Z.ai: GLM 5.3缓存读$0.07 → $0.10+44%
- Z.ai: GLM 5.3输出价$1.40 → $2.77+98%
- Z.ai: GLM 5.3输入价$0.07 → $0.10+49%
- Z.ai: GLM 5.3 Flash最大输出944K → 131K
- Z.ai: GLM 5.3 Flash缓存读$0.03 → $0.02-34%
- Z.ai: GLM 5.3 Flash输出价$0.50 → $0.13-74%
- MiniMax: MiniMax M3.1 Flash缓存读$0.01 → $0.01-50%
- MiniMax: MiniMax M3.1 Flash输出价$0.60 → $0.30-50%
- MiniMax: MiniMax M3.1 Flash输入价$0.15 → $0.07-50%
- Particle AI: Particle-1输出价$1.00 → $0.50-50%
2026-10-29
- Z.ai: GLM 4.6缓存读$0.11 → $0.10-9%
- Z.ai: GLM 4.6输出价$2.20 → $2.00-9%
- Z.ai: GLM 4.6输入价$0.55 → $0.50-9%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)最大输出384K → 393K
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.07 → $0.14+82%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.78 → $3.20+80%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.89 → $1.60+80%
- Z.ai: GLM 5.2缓存读$0.03 → $0.01-85%
- Z.ai: GLM 5.2输出价$5.40 → $7.00+30%
- Z.ai: GLM 5.2输入价$0.03 → $0.01-82%
- MoonshotAI: Kimi K3缓存读$0.18 → $0.21+15%
- MoonshotAI: Kimi K3输出价$20.00 → $10.80-46%
- MoonshotAI: Kimi K3输入价$0.19 → $0.24+24%
- Auto Router (Beta)最大输出— → 66K
- Qwen: Qwen3.8 27B最大输出236K → 262K
- Qwen: Qwen3.8 27B缓存读$0.09 → $0.05-44%
- Z.ai: GLM 5.3 Flash最大输出131K → 944K
- Z.ai: GLM 5.3 Flash输入价$0.03 → $0.15+335%
- DeepSeek: DeepSeek V4.1 Flash输入价$0.02 → $0.04+125%
- Mistral: Mistral Large 4缓存读$0.14 → $0.07-50%
- Mistral: Mistral Large 4输出价$4.18 → $2.10-50%
- Mistral: Mistral Large 4输入价$1.36 → $0.70-49%
2026-10-28
- Google: Gemma 3 4B缓存读$0.02 → —
- DeepSeek: DeepSeek V3.1最大输出33K → 147K
- Z.ai: GLM 4.6V缓存读$0.06 → $0.05-9%
- Qwen: Qwen3.5-27B缓存读$0.27 → —
- NVIDIA: Nemotron 3 Super最大输出236K → 262K
- NVIDIA: Nemotron 3 Super上下文窗口262K → 1.05M
- Z.ai: GLM 5.1最大输出128K → 131K
- Z.ai: GLM 5.1缓存读$0.18 → $0.26+45%
- Z.ai: GLM 5.1输出价$3.04 → $4.40+45%
- Z.ai: GLM 5.1输入价$0.97 → $1.40+45%
- Xiaomi: MiMo-V2.5缓存读$0.00 → $0.08+2977%
- Xiaomi: MiMo-V2.5输出价$0.24 → $2.00+740%
- Xiaomi: MiMo-V2.5-Pro缓存读$0.00 → $0.10+3329%
- DeepSeek: DeepSeek V4 Flash 0423最大输出131K → 944K
- DeepSeek: DeepSeek V4 Flash 0423缓存读$0.01 → $0.02+92%
- DeepSeek: DeepSeek V4 Flash 0423输出价$0.09 → $0.18+92%
- DeepSeek: DeepSeek V4 Flash 0423输入价$0.05 → $0.09+92%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)缓存读$0.07 → $0.07+2%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输出价$1.75 → $1.78+2%
- DeepSeek V4 Pro 0423(旧检查点·第三方美元报价)输入价$0.88 → $0.89+2%
- Qwen: Qwen3.6 27B最大输出82K → 66K
- Qwen: Qwen3.6 27B输出价$3.20 → $2.70-16%
- Qwen: Qwen3.6 27B输入价$0.32 → $0.45+41%
- Qwen: Qwen3.6 35B A3B缓存读$0.05 → $0.03-50%
- Qwen: Qwen3.6 35B A3B输出价$0.95 → $0.90-5%
- Z.ai: GLM 5.2输出价$7.00 → $5.40-23%
- MoonshotAI: Kimi K3缓存读$0.17 → $0.18+8%
- MoonshotAI: Kimi K3输入价$0.18 → $0.19+7%
- DeepSeek: DeepSeek V4 Flash 0731缓存读$0.00 → $0.02+1600%
- Z.ai: GLM 5.3 Flash最大输出944K → 131K
- Z.ai: GLM 5.3 Flash输入价$0.15 → $0.03-77%
- IBM: Granite 4.2 8B输出价$0.25 → $0.15-40%
- DeepSeek: DeepSeek V4.1 Flash最大输出131K → 944K
- DeepSeek: DeepSeek V4.1 Flash缓存读$0.00 → $0.01+43%
- DeepSeek: DeepSeek V4.1 Flash输出价$0.42 → $0.50+19%
- inclusionAI: Ling 3.1 Flash最大输出236K → 262K
- inclusionAI: Ling 3.1 Flash上下文窗口262K → 1M
- Meta: Muse Spark Open最大输出944K → 131K
2026-10-27
- Google: Gemma 3 4B最大输出118K → 16K
- Google: Gemma 3 4B输出价$0.05 → $0.10+100%
- Google: Gemma 3 4B输入价$0.04 → $0.05+25%
- Qwen: Qwen3 30B A3B Instruct 2507缓存读— → $0.10
- Nous: Hermes 4 405B缓存读— → $0.20
- Z.ai: GLM 4.6V缓存读$0.05 → $0.06+10%
- NVIDIA: Nemotron 3 Nano 30B A3B缓存读— → $0.01
- Qwen: Qwen3.5-27B最大输出66K → 82K
- Qwen: Qwen3.5-27B输出价$2.16 → $2.60+20%
- Qwen: Qwen3.5-27B输入价$0.27 → $0.26-4%
- Z.ai: GLM 5.1最大输出131K → 128K
- Z.ai: GLM 5.1缓存读$0.26 → $0.18-31%
- Z.ai: GLM 5.1输出价$4.40 → $3.04-31%
价格数据参考 OpenRouter 及各厂商公开页面整理,以 美元 / 百万 tokens 计,可能滞后或有误,请以官方页面为准。不构成任何采购或投资建议。