陈天桥代季峰打响2026大模型第一枪：30B参数跑出1T性能 – 量子位
Quick Answer
MiroThinker 1.5, developed by Chen Tianqiao, showcases a remarkable performance of 1T with just 30B parameters, outperforming competitors like Kimi-K2-Thinking in benchmarks while maintaining a lower inference cost of $0.07 per call.
Quick Take
MiroThinker 1.5, developed by Chen Tianqiao, showcases a remarkable performance of 1T with just 30B parameters, outperforming competitors like Kimi-K2-Thinking in benchmarks while maintaining a lower inference cost of $0.07 per call. This model challenges the notion that larger parameter sizes equate to better performance.
Key Points
- MiroThinker 1.5 scored 39.2% on -Text and 71.5% on BrowseComp-ZH.
- It has only 30B parameters, compared to competitors with 235B.
- Inference cost for MiroThinker 1.5 is $0.07 per call, 1/20 of Kimi's cost.
- MiroThinker 1.5 outperformed Kimi-K2-Thinking by 4.5% in BrowseComp-ZH.
- The model emphasizes efficiency over sheer parameter size.
Article Excerpt
From source RSS / original summary< img id="wx\_img" src="https://www. qbitai. com/wp-content/uploads/imgs/qbitai-logo-1. png" width="400" height="400">. # 陈天桥代季峰打响2026大模型第一枪：30B参数跑出1T性能. *鹭羽* 2026-01-06 14:28:58 来源：量子位. 正式发布其自研的旗舰版搜索智能体模型——**MiroThinker 1. 5**，堪称智能体模型领域的最强小钢炮。. 在面对GPT-5-High、Gemini-3-Pro、DeepSeek-V3. 2等一系列国内外顶尖模型，MiroThinker 1. 5在四项基准测试中的表现都毫不逊色：. * -Text（人类终极测试）：39. 2%. * BrowseComp-ZH（BrowseComp的中文适配版本）：71. 5%. * GAIA-Val-165（GAIA基准测试验证集）：80. 8%. 但要知道MiroThinker 1. 5的参数规模只有它们的**1/30**，仅30B和235B大小。.
换句话说，在全行业大模型都在卷参数规模和上下文长度时，MiroThinker 1. 5直接用高智效比跑出了相近甚至更好的效果，原因就在于它抓住了这股“巧劲”：. 比如将MiroThinker-v1. 5-30B和1T参数的Kimi-K2-Thinking对比，不仅在BrowseComp-ZH测试中实现了4. 5%的性能超越，在推理成本上，MiroThinker 1. 5的单条调用成本更是低至$0. 07，只有Kimi的1/20。. 不止如此，MiroThinker 1. 5的推理速度也显著优于Kimi-K2-Thinking，足以见得**“大”不等于“强”**，叠参数也并非大模型的唯一通解。. 而负责操刀这款模型的正是**MiroMin
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from WebSearch (Tavily)
See more →WSJ: OpenAI is considering deep price reductions as competition ...
OpenAI is contemplating significant price cuts in response to competitive pressure from Anthropic, particularly due to the success of Claude Code in developer and coding workflows. This shift could affect pricing strategies in the AI market as companies vie for dominance in coding solutions.