小米9月22日發佈開源全模態模型MiMo-V2.6系列（Pro/Flash），基於可驗證任務擴展強化學習，探索遞歸自我改進。Pro版在Artificial Analysis綜合智能指數獲46分，超越Kimi K3與Qwen3.8Max，成最強開源模型；多數Agent基準比肩Claude Opus5與GPT-5.6Sol，但與最強閉源仍有差距。兩款模型Live RL訓練不足6天，成本約85萬和262萬元。
DeepSeek近期動作密集：本週低調發布V4.1 Flash，雖未調整大版本號，但底層架構升級幅度已達V5級別；後續還將推出全新大模型Code2.0及V4.1 Pro系列，其架構換代與新品路線引發行業關注。
深度求索發佈DeepSeek V4.1 Flash，爲全新結構系列最小模型，具備原生多模態視覺理解。新結構旨在提升能力上限、推理速度與吞吐，並可擴展至更大參數。該模型爲552B參數MoE，採用Causal-Encoder-Decoder不對稱結構：輸入激活僅8B，輸出激活16B，HBM需求降至1/4，顯著降低成本。
DeepSeek宣佈自2026年9月10日12時起調整Flash系列模型價格：空閒時段輸入緩存命中、未命中和輸出每百萬Token分別爲0.02元、1元、4元，高峯時段翻倍；三檔價格全線下調，緩存命中降幅達60%。涉及deepseek-v4-flash及deepseek-v4-flash-vision-exp兩款模型。
Google
$0.49
輸入tokens/百萬
$2.1
輸出tokens/百萬
1k
上下文長度
Openai
$2.8
$11.2
Xai
$1.4
$3.5
2k
$7.7
$30.8
200
-
Anthropic
$105
$525
$0.7
$7
$35
$17.5
$21
Alibaba
$1
$10
256
$6
$24
$2
$20
$4
$16
Baidu
128
Bytedance
$1.2
$3.6
4
$8
$240
52
%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
VIP會員
