hermes - 💡(How to fix) Fix fix: update default Gemini model from gemini-2.5-flash to gemini-3.5-flash (shutdown Oct 16 2026)

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…

gemini-2.5-flash is hardcoded as the default Gemini model in ~15 places across the codebase. Google has published a deprecation schedule with a shutdown date of October 16, 2026. The recommended replacement is gemini-3.5-flash (stable GA, 1M context, 65K output).

After the shutdown date, every user who has not manually overridden their model config will start getting API errors with no clear explanation.

Root Cause

gemini-2.5-flash is hardcoded as the default Gemini model in ~15 places across the codebase. Google has published a deprecation schedule with a shutdown date of October 16, 2026. The recommended replacement is gemini-3.5-flash (stable GA, 1M context, 65K output).

After the shutdown date, every user who has not manually overridden their model config will start getting API errors with no clear explanation.

RAW_BUFFERClick to expand / collapse

Summary

gemini-2.5-flash is hardcoded as the default Gemini model in ~15 places across the codebase. Google has published a deprecation schedule with a shutdown date of October 16, 2026. The recommended replacement is gemini-3.5-flash (stable GA, 1M context, 65K output).

After the shutdown date, every user who has not manually overridden their model config will start getting API errors with no clear explanation.

Affected locations

Functional defaults (break on shutdown):

  • agent/gemini_native_adapter.pyprobe_gemini_tier default model param and _create_chat_completion default model param
  • agent/gemini_cloudcode_adapter.py — same default model param
  • agent/gemini_native_adapter.py_FREE_TIER_GUIDANCE text and free-tier rate comment still reference gemini-2.5-flash
  • hermes_cli/main.py — doctor warning mentions gemini-2.5-flash rate

Aux model defaults (summarization, vision, web extraction):

  • agent/auxiliary_client.pygemini-3-flash-preview aux model default (also outdated)
  • plugins/model-providers/gemini/__init__.pydefault_aux_model
  • plugins/model-providers/openrouter/__init__.py — OpenRouter aux default
  • plugins/model-providers/kilocode/__init__.py — same
  • plugins/memory/hindsight/__init__.py — hardcoded gemini-2.5-flash

Model picker / setup wizard:

  • hermes_cli/models.py — featured model lists
  • hermes_cli/setup.py — setup wizard model choices

Comments / docstrings (low priority):

  • hermes_cli/config.py, hermes_cli/goals.py, tools/web_tools.py, tools/vision_tools.py, trajectory_compressor.py, datagen-config-examples/trajectory_compression.yaml

Expected behaviour

gemini-3.5-flash (and google/gemini-3.5-flash for OpenRouter paths) should be the default everywhere gemini-2.5-flash or gemini-3-flash-preview currently appears as a hardcoded default.

Side effects

Likely also resolves #26623 (hermes doctor falsely reports a valid Gemini key as invalid) since probe_gemini_tier would probe against the non-deprecated model.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING