Ricursive Intelligence released the open-source model Ripple-65B Preview, stating that it scores 80.5% on the Terminal-Bench 2.1 programming test, and the model weights use the Apache 2.0 license.
Cognition launches coding model SWE-2 Flash, claiming significantly improved speed and cost over the previous generation, leading in some tests but with lower Terminal-Bench performance. The model is now available on Devin Desktop and CLI.
Paradigm’s Chief Technology Officer stated that Solar will be released this week EVM code generation feature, which is over solc times faster than 100 and can generate bytecode with higher Gas efficiency.
Inception released Mercury 3, stating that its parameter count is more than 40 times larger than the first generation, and it is now available on API、OpenRouter, Baseten, and Vercel. There is a significant discrepancy between the output speed reported in official benchmarks and that shown by third-party platforms.
JetBrains reportedly open-sourced a 12000000000-parameter coding model Mellum2.1, and released its weights and GGUF quantized version. The report cited its self-reported results, saying that the model scored higher than LiveCodeBench v6 on Qwen3.5-9B。
InclusionAI releases Ling-3.1-flash, with a clear increase in parameter scale and per-token activation scale compared with the previous generation. The official demonstration showed it completing long-duration programming tasks, with plans to make context and an open-source model available later.