Expired50% confidenceFactTime unknown
In the LM Arena benchmark, 3.5 Flash surpassed the previous-generation flagship 3.1 Pro in areas such as SVG generation, 3D programming, and animation processing
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Gemini 3.5 Pro Leak Analysis: Coding Matches GPT 5.5, Spark Agent Sparks Privacy Controversy
bilibiliAI全文总结
Related Claims
UnverifiedStep 3.7 Flash 在 V* 基准测试中达到 95.3 分73% similarUnverified在Three.js可变速齿轮机械动画测试中，DeepSeek V4 Pro的效果远超此前V4 Flash版本，与Kimi K3基本持平70% similarUnverifiedIn the FPS game generation test, V4 Flash generated 988 lines of code while V4 Pro generated 1,456 lines of code.65% similarUnverifiedThe Flash series from 1.5 to 2.0, 2.5, and then 3.0 has seen capability improvements far exceeding expectations.63% similarUnverifiedThe STEP3.7 Flash large model released by StepFun ranked first on the AA Benchmark (Artificial Analysis), covering three dimensions: speed, cost-effectiveness, and end-to-end multimodal performance.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/55651API
curl https://kongchang.com/api/v1/knowledge/claims/55651MCP
get_claim(id=55651)