22939a2580
Revert "feat(llm/google): forward cached_content into generate_content ( #4889 )"
...
This reverts commit 5a745a8502640360e9b7
2026-05-23 12:03:38 -07:00
df063f9068
feat(llm/google): forward cached_content into generate_content
...
Adds a cached_content field on ChatGoogle (and a per-call kwarg on
ainvoke) that gets threaded into the GenerateContentConfigDict before
each generate_content call. This lets callers point Gemini at an
explicit CachedContent resource (e.g. "cachedContents/abc123") instead
of relying on implicit caching, which is constrained by the ~5-minute
TTL window.
Token accounting already pulls cached_content_token_count from
response.usage_metadata into ChatInvokeUsage.prompt_cached_tokens, so
the savings show up in usage stats without further work.
Cache creation itself (client.caches.create) is left to the caller —
this PR only adds the forwarding hook so explicit caching becomes
opt-in usable. A follow-up can wire agent-side cache lifecycle if
useful.
2026-05-22 17:54:04 -07:00
5ef0d9dbac
feat(llm/google): add gemini-3.1-pro-preview to verified models
...
Routes through the existing Gemini 3 Pro thinking branch by extending
is_gemini_3_pro to match 'gemini-3.1-pro' alongside 'gemini-3-pro'.
2026-05-22 11:48:10 -07:00
9e8dfdd18f
fix(llm/google): route gemini-3.1-flash through gemini-3 flash thinking branch
2026-05-22 10:50:35 -07:00
5c54ef4197
chore(llm): rename flash-lite recommendation to gemini-3.1-flash-lite
2026-05-22 10:49:12 -07:00
b00c41b66b
chore(llm): recommend gemini-3-flash-preview in examples and tests
...
- add gemini-3-flash-preview-lite to VerifiedGeminiModels and token mappings
- list both gemini-3-flash-preview[-lite] alongside existing -latest aliases in CLOUD.md and api-v2.md SupportedLLMs
- swap example code and recommendations (examples/, AGENTS.md, quickstart.md, bug report placeholder) to gemini-3-flash-preview / -lite
- update Google CI test + evaluate_tasks judge LLM to gemini-3-flash-preview / -lite
2026-05-22 10:47:58 -07:00
115199d2ba
fix: handle types.HttpOptionsDict better, add more tests
2026-05-22 14:30:40 +08:00
2eb3e3cded
feat: add client header to GoogleChat
...
Added header per integration
[guidelines](https://ai.google.dev/gemini-api/docs/partner-integration ).
2026-05-22 14:20:24 +08:00
90783e69b4
fix: set default temperature to 1.0 for gemini-3 models
2026-03-23 14:13:08 +08:00
6d54f521cf
set default gemini thinking to auto
2026-01-26 12:44:51 -08:00
bb3b1e8cf5
added minimal model for gemini 3 flash
2026-01-23 15:11:13 -08:00
041d03a93f
feat: updated the thinking budget
2026-01-23 14:42:59 -08:00
5b06f4edd9
set default thinking to 0 for gemini-3 models
2026-01-23 13:44:29 -08:00
221d85744e
fix: use correct MIME type for images in Google serializer
2025-12-21 09:45:09 -08:00
eb2382c7c4
added gemini 3 flash preview model
2025-12-18 11:25:23 -08:00
dc0db88215
kv caching for BU agents
2025-12-15 18:23:22 -08:00
49882fb086
added gemini-3-pro-preview to verified models
2025-12-09 13:38:34 -08:00
4d87e49e1f
Merge branch 'main' into mert/structured_output
2025-11-25 13:43:55 -08:00
ed7cfa4ab2
fix title problem in schema as well as enum problems
2025-11-25 13:02:39 -08:00
84be050ed7
exponential retries improved
2025-11-25 11:18:03 -08:00
af649d357a
Retry 503 errors
2025-10-29 23:04:58 -07:00
b27d7fe9d2
Retry 503 errors
2025-10-29 23:02:06 -07:00
9715c4e764
Linter
2025-10-29 17:14:45 -07:00
577c1ff148
Retry google on 403 error.
2025-10-29 16:51:32 -07:00
9caa248316
Merge code7 branch changes
2025-10-21 18:55:54 -07:00
98995449c2
reduce max_output_tokens gemini
2025-10-14 14:31:57 -07:00
3aaafa7bf8
Remove required from prompt
2025-10-05 20:30:36 -07:00
a3a811e3be
Remove create_gemini_optimized_schema
2025-10-05 19:03:10 -07:00
968414b982
Compress gemini schema
2025-10-05 12:48:25 -07:00
a8e6af254e
Switch all to jpeg
2025-10-03 21:32:47 -07:00
781f948bb8
change gemini 2.5 flash to gemini latest
2025-09-26 15:54:19 -07:00
6c7b5cd622
fix-gemini-thinking-budget-for-latest-lite-version
2025-09-26 12:09:03 -07:00
6d77cc6bb9
fix-gemini-thinking-budget-for-latest-version
2025-09-26 11:53:58 -07:00
db27b32b04
fix gemini client
2025-09-26 09:50:01 -07:00
cf0ad68b4c
Merge branch 'speed-ups3' of https://github.com/browser-use/browser-use into speed-ups3
2025-09-22 18:37:59 -07:00
343b766b09
Comment
2025-09-22 18:37:57 -07:00
c8e00e4511
Update browser_use/llm/google/chat.py
...
Co-authored-by: cubic-dev-ai[bot] <191113872+cubic-dev-ai[bot]@users.noreply.github.com>
2025-09-22 18:35:26 -07:00
5d749793ec
Refactor model timeout logic and enhance ModelProviderError initialization
2025-09-22 16:02:03 -07:00
d5ba9008b7
Update thinking_budget default value to 0 for flash and flash-lite models
2025-09-22 14:50:24 -07:00
aa56461dbc
Update thinking_budget default value to -1 for specific model
2025-09-20 07:23:17 -07:00
545dc86295
Include thinking_budget 0
2025-09-19 23:40:32 -07:00
13a83b7321
Include thinking_budget 600
2025-09-19 23:22:56 -07:00
382bbd45fc
Include thinking_budget -1
2025-09-19 23:14:28 -07:00
653fa4206f
Automaticlly set reasoning budget to 0 for 2.5 flash, because we do it with thinking for memory
2025-09-19 22:21:37 -07:00
fbcde93cad
Force no thinking
2025-09-19 19:40:33 -07:00
af1fe232f1
Forse thinking -1
2025-09-19 19:39:18 -07:00
7e2d4de32c
Timing
2025-09-19 19:36:53 -07:00
0aa3163693
Include response in parsing
2025-09-19 19:21:48 -07:00
b72dea7972
Improve error logging in ChatGoogle class by including response details in the error message when no response is received from the model.
2025-09-19 19:15:44 -07:00
819aa77838
Update thinking_budget handling in ChatGoogle class to use dynamic value
2025-09-19 19:14:21 -07:00
MagMueller