[Feedback/Feature] Dynamic thinking budget floor & explicit sampling controls for multimodal reasoning in Gemini Flash

[adri22235]

Summary

We have conducted extensive empirical testing comparing gemini-3.8-flash across reasoning effort levels (Medium vs. High) on complex multimodal tasks (fine-grained OCR on compressed video, phonetic audio matching under background noise, and multi-step regulatory decision-making).

Observed Behavior

1. Truncated Test-Time Compute in Medium:

   • On ambiguous OCR tasks (e.g., distinguishing digits 5 vs. 6 on compressed 240p/720p video feeds or legacy VGA BIOS fonts), Medium thinking effort frequently locks in early on statistical priors (e.g., hallucinating version 1.50 instead of 1.60 despite the text being explicitly visible).
   • Once a false premise is established in the initial reasoning tokens, the output generation locks onto the hallucinated token with ~99% confidence, bypassing default sampling filters.

2. Impact of Fixed Sampling (temperature: 1.0, top_p: 0.95):

   • While High reasoning effort completely eliminates these hallucinations due to sufficient compute for cross-frame verification and hypothesis pruning, Medium suffers from early budget exhaustion under temperature: 1.0.

3. In High Mode:

   • Accuracy reaches ~100% across fine OCR, license plate reconstruction (2291 KWJ), noisy audio parsing, and complex DGT traffic regulations.

Proposed Suggestions for the SDK / Backend:

1. Dynamic Thinking Floor for Multimodal Inputs:
   • Implement an adaptive minimum token budget for Medium when video or dense audio tokens are present in the payload, ensuring at least one full verification/falsification pass.
2. Explicit SDK & UI Controls for top_p and thinking_budget:
   • Keep top_p and explicit thinking_budget integer controls accessible in types.GenerateContentConfig and clearly surfaced in AI Studio documentation for power users who need deterministic calibration.
3. Improved FinishReason.MALFORMED_RESPONSE Handling:
   • Enhance client-side retry or fallback handling when reasoning outputs encounter malformed structures during rapid token generation.

Environment

• SDK: googleapis/python-genai
• OS: Debian 12 (Bookworm) / Linux
• Model: gemini-3.8-flash / gemini-3.7-flash

[Venkaiahbabuneelam]

Hi @adri22235, Thank you for bringing this issue to our attention. we are filing this as a Feature Request, to enhance multimodal reasoning calibration and SDK control surface for gemini-3.8-flash / gemini-3.7-flash.

Thanks