Fix: Update deprecated Gemini model and error logging for silent bot response failures.

[syedbarkath980]

Problem

Vs-bot debate responses and judging were failing silently, returning generic personality-flavored fallback messages instead of real AI-generated responses — even with a valid and correctly configured Gemini API key.

There were two separate issues contributing to the problem.

---

Root Cause

1. Gemini errors were silently swallowed

GenerateBotResponse discarded the actual Gemini error and returned a hardcoded, in-character fallback message instead.

For example, when the Rookie Rick bot's Gemini request failed, the user would receive:

Uh, wait a sec! Like, I totally blanked out, you know? My bad, kinda like that time at Cousin Joey's BBQ!

This fallback is defined per bot in personalityErrorResponse:

case "Rookie Rick":
    return fmt.Sprintf(
        "%s Like, I totally blanked out, you know? My bad, kinda like that time at Cousin Joey's BBQ!",
        catchphrase,
    )

The issue was inconsistent with JudgeDebate in the same file, which already logs Gemini errors:

// JudgeDebate — logs the error
if err != nil {
    log.Printf("Gemini error: %v", err)
}

Whereas GenerateBotResponse previously did this:

// GenerateBotResponse — error was not logged
if err != nil {
    return personalityErrorResponse(botName, "A glitch in my logic, there is.")
}

As a result, the underlying Gemini failure was completely invisible in the server logs.

Every request still returned a 200 OK with a plausible-looking, in-character response, making the failure appear to be normal application behavior.

The actual error only became visible after temporary debug logging was added to GenerateBotResponse.

---

2. Deprecated Gemini model

Once the underlying error was exposed, the actual failure was identified as a 404 from Google:

Error 404: This model models/gemini-2.5-flash is no longer available to
new users. Please update your code to use models/gemini-3.6-flash for
the latest features and improvements.

The project was using the hardcoded model:

gemini-2.5-flash

This model is no longer available to new Gemini API keys, meaning new contributors setting up the project with a freshly created API key could encounter the same failure.

---

Fix

Two changes were made:

Error logging

Added error logging in case of failure for Gemini API errors in GenerateBotResponse

This brings GenerateBotResponse in line with the existing error-handling behavior in JudgeDebate.

Gemini model update

Updated the default Gemini model:

- const defaultGeminiModel = "gemini-2.5-flash"
+ const defaultGeminiModel = "gemini-3.6-flash"

---

NOTE: Needs Clarification for the Existing users with 2.5 version:

Google's error message specifically states that gemini-2.5-flash is "no longer available to new users."

This may imply that existing API keys or deployments could still be able to use the older model. I wasn't able to verify this independently because testing was performed with a newly created Gemini API key.

Therefore, if an existing deployment is currently relying on gemini-2.5-flash, this model change could potentially affect it as well.

This is being called out explicitly rather than assuming that existing users are unaffected.

---

Before

Vs-bot returned generic personality-specific fallback messages instead of actual debate responses:

"Uh, wait a sec! Like, I totally blanked out, you know?
My bad, kinda like that time at Cousin Joey's BBQ!"

The Gemini failure was also not visible in the server logs.

improper.ai.response.before-fix.mp4

---

After

Vs-bot successfully returns contextual, AI-generated debate responses using the updated Gemini model.

improper.ai.response.after-fix.mp4

Summary by CodeRabbit

• Bug Fixes
  • Improved error visibility when AI-generated debate responses fail.
  • Updated the default AI model used to generate responses.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 1671bd34-34b3-4fc1-816c-93632643c4b3

📥 Commits

Reviewing files that changed from the base of the PR and between dfd6b75 and 8516d06.

📒 Files selected for processing (2)

• backend/services/debatevsbot.go
• backend/services/gemini.go

Included review availability: Your plan provides up to 1 included review per hour; 0 remain after this review.

---

📝 Walkthrough

Walkthrough

The default Gemini model changed to gemini-3.6-flash. GenerateBotResponse now logs Gemini generation errors before returning its existing fallback response.

Changes

Gemini response handling

Layer / File(s)	Summary
Model selection and error logging backend/services/gemini.go, backend/services/debatevsbot.go	The default model uses gemini-3.6-flash. GenerateBotResponse logs generation errors with log.Printf before returning the personality-specific fallback response.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Merge Risk: ⚪ Minimal · up to 8516d

This localized change updates the default Gemini model and improves visibility into failed responses; no actionable merge-blocking risk remains beyond normal checks and review.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 2 functions across 2 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies both main changes: updating the deprecated Gemini model and adding error logging for bot response failures.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[Ri1tik]

@syedbarkath980 Good Catch but I have never seen such msg till now.

[syedbarkath980]

No, that message was shown AFTER i added a log line: log.Printf("❌ Gemini error in GenerateBotResponse: %v", err) in GenerateBotResponse function which was not present before i made the change. It is added to catch and show the error thrown by gemini in case of any issue caused by the gemini api.