Gemini last models: temperature, top_p, and top_k are deprecated and ignored
ai.google.dev
July 21, 2026
10 min read
🔥🔥🔥🔥🔥
45/100
Summary
Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now generally available for production use. Gemini 3.6 Flash offers improved performance in complex tasks with reduced token usage and lower pricing compared to 3.5 Flash, while Gemini 3.5 Flash-Lite is the fastest and most cost-effective model in the 3.5 family, outperforming previous Flash-Lite generations.
Key Takeaways
- Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now generally available for production use.
- Gemini 3.6 Flash offers improved performance in complex tasks with reduced token usage and lower pricing compared to Gemini 3.5 Flash.
- Gemini 3.5 Flash-Lite is the fastest and most cost-effective model in the 3.5 family, outperforming previous generations in high-performance execution.
- Both models support a context window of 1 million tokens and a maximum output of 64,000 tokens, along with integrated tools for enhanced functionality.
Community Sentiment
Positives
- Some users appreciate the potential for improved determinism with system instructions, suggesting a shift towards more tailored AI interactions.
Concerns
- Commenters are concerned that the removal of top_k and top_p will lead to less control over output quality, with one predicting it will be worse than using top_k=1.
- There's skepticism about the stability of Gemini 4, with users pointing out that it becomes unstable when not using the 'official' sampler settings.
