Google Set to Launch Gemini Model That Could Beat Rivals in Coding

Google is reportedly preparing to launch a new Gemini artificial intelligence model that could strengthen its position in the increasingly competitive AI coding market. Internally known by the codename “Skimaki”, the model is expected to be called Gemini 3.8 Flash and could be released soon. The model is being developed with a particular focus on coding capabilities, an area where rival AI systems from Anthropic and OpenAI have gained significant attention among developers and businesses.

The upcoming model is reportedly being evaluated internally through Google’s coding platform, where some engineers have preferred its performance over Anthropic’s Opus model in head-to-head testing. However, Gemini 3.8 Flash is not expected to be positioned as a new flagship frontier model. Google’s Flash series is designed to provide faster responses while requiring fewer computing resources than larger Pro models, making the models more suitable for rapid development, testing and deployment across different applications.

The release comes as Google works to address challenges surrounding its broader Gemini strategy. The company has faced delays involving its next-generation Pro model after internal versions reportedly failed to deliver sufficient improvements on coding benchmarks. Google has also been rapidly updating its Flash lineup, with newer versions appearing within relatively short intervals. This approach allows its teams to experiment with different model architectures and training techniques without committing the same level of computing resources required for larger flagship systems.

Also Read: Google DeepMind Team Offers Applicants a Way Around AI Filters

Coding has become an increasingly important battleground in the AI industry as companies develop systems capable of writing, debugging and modifying software with limited human intervention. Google has consequently increased its focus on coding performance and reinforcement learning, while also bringing in experienced AI researchers to strengthen its post-training efforts. The company is simultaneously developing Gemini 4, which has reportedly performed well in early pretraining evaluations but remains under development as researchers continue working on its post-training process.

Google is facing competition from several directions as it attempts to improve its position in AI. Anthropic and OpenAI have established strong positions in coding-focused AI tools, while other technology companies are also developing competing coding agents and models. Google’s Gemini platform has nevertheless continued to attract a large user base, but the company faces pressure to translate that adoption into stronger enterprise usage, where AI-powered coding assistants and software-development agents are becoming increasingly important.

The reported Gemini 3.8 Flash launch could therefore be an important step in Google’s effort to regain momentum in AI development. A stronger coding-focused model could help the company compete more effectively in a rapidly evolving market while providing developers with a faster and potentially more cost-efficient option. The model’s actual standing against competing systems, however, will depend on independent testing and real-world performance after its wider release rather than internal evaluations alone.

Also Read: JP Nadda Assures Withdrawal Of FIRs After CJP Calls Off September 5 Protest