Will Li
e2b13a6518
code review + speed fix
2025-07-29 20:02:09 -07:00
Matt Rubens
8a3dcfb593
Update the max-token calculation in model-params to use the shared logic ( #5720 )
2025-07-15 02:20:22 -04:00
Matt Rubens
5762964b56
Add Kimi K2 model and better support ( #5717 )
2025-07-15 00:31:18 -04:00
Hannes Rudolph
f5dfbf1fc2
feat(claude-code): add configurable max output tokens setting ( #5610 )
...
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Daniel <57051444+daniel-lxs@users.noreply.github.com>
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
2025-07-11 22:31:50 -04:00
Shariq Riaz
9faf28acac
fix: use actual max_completion_tokens from OpenRouter API ( #5240 )
...
- Update parseOpenRouterModel to always use actual max_completion_tokens from OpenRouter API
- Remove artificial restriction that only reasoning budget and Anthropic models get their actual max tokens
- Fall back to 20% of context window when max_completion_tokens is null
- Update getModelMaxOutputTokens to use same fallback logic for consistency
- Update tests to reflect new behavior
- Fixes issue where reserved tokens showed ~209k instead of actual model limits (e.g. GPT-4o: 16,384)
2025-07-06 04:15:05 -04:00
Brad Davis
37ed013157
fix context length for lmstudio and ollama ( #2462 ) ( #4314 )
...
Co-authored-by: Daniel Riccio <ricciodaniel98@gmail.com>
2025-06-20 22:53:04 -04:00
Chris Estreich
9e3b0fd870
Move provider data into @roo-code/types ( #4153 )
2025-05-30 14:54:48 -07:00
zeo
0ef52c8cd7
Add DeepSeek R1 0528 to the Chutes provider ( #4105 )
...
New model for the Chutes provider, surpasses the old DeepSeek R1
and is free so it is also set to be the default model now.
2025-05-29 06:41:05 -04:00
slytechnical
73d162305a
Added a hardcoded list of computer use models for litellm as a fallba… ( #4052 )
...
Added a hardcoded list of computer use models for litellm as a fallback for older litellm versions
2025-05-27 14:58:20 -05:00
Chris Estreich
e66136f1aa
Add a new @roo-code/types package and use it everywhere ( #3912 )
2025-05-26 12:06:45 -07:00
Peter Dave Hello
6ebc95d2a8
Update xaiModels and xaiDefaultModelId in src/shared/api.ts ( #3957 )
...
* Add non-beta versions of `grok-3` models to the `xaiModels` object
* Change the default `xaiDefaultModelId` from `grok-3-beta` to `grok-3`
2025-05-25 09:10:46 -04:00
Chris Estreich
77a5b6c06d
Fix reasoning budget for Gemini 2.5 Flash on OpenRouter ( #3945 )
...
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
2025-05-24 19:46:42 -04:00
Matt Rubens
4abadd877b
Default to Sonnet 4 ( #3928 )
2025-05-24 13:06:25 -04:00
Shariq Riaz
ff50a47b5b
feat: enable prompt caching for gemini-2.5-flash-preview-05-20 ( #3907 )
...
* feat: enable prompt caching for gemini-2.5-flash-preview-05-20
* feat: Add cache pricing for gemini-2.5-flash-preview-05-20 models
2025-05-24 07:38:45 -04:00
Shariq Riaz
20c7453425
Remove claude-3.7-sonnet and claude-3.7-sonnet:thinking from vscodelm ( #3895 )
...
* Remove claude-3.7-sonnet and claude-3.7-sonnet:thinking from vscodelm provider in api.ts (not supported via API)
* Add static blacklist for unsupported VS Code Language Model IDs
---------
Co-authored-by: Ruakij <ruakij@ruekov.eu>
2025-05-23 17:28:42 -04:00
slytechnical
1791bb9053
Ability to refresh LiteLLM models list (take 2) ( #3852 )
...
* Litellm models can now be refreshed
* Fix no-case-declarations lint issue and put back missing autoCondenseContextPercent webviewMessageHandler case
* Add tests for litellm config changes
* replace hardcoded keys with constants
2025-05-23 16:19:40 -05:00
Chris Estreich
872902754a
Fix max tokens in the task header ( #3893 )
2025-05-23 02:32:21 -07:00
Chris Estreich
fa1e7b415c
Re-work support for reasoning models, including "hybrid" reasoning models ( #3870 )
2025-05-23 01:40:17 -07:00
Matt Rubens
c52d6d52ab
Fix openrouter names for caching and computer use ( #3868 )
...
Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-05-22 17:07:56 -04:00
Matt Rubens
fb3f6d0378
Fix Claude 4 models in Vertex ( #3849 )
2025-05-22 14:43:31 -04:00
Shariq Riaz
8a8fadd407
Add Claude Sonnet 4 and Claude Opus 4 models with thinking variants ( #3844 )
...
Co-authored-by: cte <cestreich@gmail.com>
2025-05-22 10:44:56 -07:00
Daniel
27e47de3aa
Add Gemini Flash 2.5 05-20 variants for the Vertex provider ( #3758 )
...
* feat(api): add gemini-2.5-flash-preview-05-20 model configuration
* feat(tests): update apiModelId to gemini-2.5-flash-preview-05-20 in ProviderSettingsManager tests in case the old version is deprecated
* chore: add changeset
* feat(api): update vertexModels to add gemini-2.5-flash-preview-05-20 variants
* chore: update changeset
2025-05-20 21:13:21 -04:00
Shariq Riaz
5ccce3a39d
feat: add gemini-2.5-flash-preview-05-20 models ( #3769 )
2025-05-20 21:08:07 -04:00
zeo
d70d0c1627
Add Qwen3 model series to the Chutes provider ( #3710 )
...
* Add Qwen3 model series to the Chutes provider
New models for the Chutes provider:
- Qwen/Qwen3-235B-A22B
- Qwen/Qwen3-32B
- Qwen/Qwen3-30B-A3B
- Qwen/Qwen3-14B
- Qwen/Qwen3-8B
* add changeset
2025-05-18 09:30:45 -07:00
Chris Estreich
20fab97b4d
Gemini implicit caching ( #3515 )
2025-05-12 14:16:27 -07:00
John Richmond
8cb5c4e9e3
Restore profile management work + fix #3434 ( #3449 )
...
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
2025-05-12 15:47:09 -04:00
Matt Rubens
730d037a12
Revert "Improve provider profile management in the external API ( #3386 )" ( #3440 )
2025-05-10 08:27:37 -04:00
Chris Estreich
c21aa230e3
Get rid of ApiConfiguration type alias, rename ApiConfigMeta to ProviderSettingsEntry ( #3380 )
2025-05-08 23:52:18 -07:00
John Richmond
0bbd3fd1df
Add LiteLLM provider ( #3242 )
...
Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-05-07 14:53:21 -04:00
zetaloop
328a25e889
Add gemini-2.5-pro-preview-05-06 model ( #3222 )
...
Add model gemini-2.5-pro-preview-05-06
2025-05-06 09:33:09 -07:00
Chris Estreich
a649a53ec6
Move remaining provider settings into separate components ( #3208 )
2025-05-05 23:09:24 -07:00
Daniel Trugman
ce8fbbdafa
Requesty provider fixes ( #3193 )
...
Co-authored-by: Chris Estreich <cestreich@gmail.com>
2025-05-05 20:49:11 -07:00
Shariq Riaz
883be32b1a
feat: Add Groq and Chutes API providers ( #3034 )
...
Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
Co-authored-by: Chris Estreich <cestreich@gmail.com>
2025-05-05 15:16:27 -07:00
Ash R
a3d8c0e3f6
feat: vertex/gemini prompt caching ( #2996 )
2025-04-28 15:11:35 -07:00
Chris Estreich
cb29e9d56f
Remove ModelInfo objects from settings ( #2939 )
2025-04-25 15:23:25 -07:00
pugazhendhi-m
7e76736e13
Updates default model for Unbound ( #2944 )
...
* Updates default model for Unbound
* Adds changeset
---------
Co-authored-by: Pugazhendhi <pugazhendhi@unboundsecurity.ai>
2025-04-25 07:36:25 -04:00
Chris Estreich
0dfbae64f3
Allow users to toggle Gemini caching on / off for OpenRouter ( #2927 )
2025-04-24 16:34:22 -07:00
Chris Estreich
a3f1a3f3ad
Gemini caching improvements ( #2925 )
2025-04-24 14:20:09 -07:00
Chris Estreich
b75379bed3
Improve OpenRouter model fetching ( #2922 )
2025-04-24 12:29:36 -07:00
Chris Estreich
ad4782b766
Add an option to enable prompt caching ( #2924 )
2025-04-24 12:28:43 -07:00
Chris Estreich
fb91836203
Revert Gemini caching, fix OR supports cache issue ( #2918 )
2025-04-24 09:15:56 -07:00
Chris Estreich
a08461a655
Gemini prompt caching ( #2827 )
2025-04-23 11:35:50 -07:00
Alfredo Medrano
3a5913ffca
Bugfix/fix vscodellm model information ( #2832 )
...
* feat: initialize VS Code Language Model client in constructor
* feat: add VS Code LLM models and configuration
* feat: integrate VS Code LLM models into API configuration normalization
* Fix tests
---------
Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
2025-04-23 03:05:47 -04:00
Nico Bihan
eeb73c3c06
Adds Gemini 2.5 Flash "thinking" model to Vertex AI Provider ( #2794 )
2025-04-19 21:30:20 -07:00
Chris Estreich
96ff9fc380
Fix pricing for Gemini 2.5 Flash (Thinking) ( #2773 )
...
* Fix pricing for Gemini 2.5 Flash (Thinking)
* Looks like it's actually $3.50
* We aren't honoring custom thinking token budgets on Vertex yet
2025-04-18 14:43:27 -07:00
Chris Estreich
5abea50cf1
Support Gemini 2.5 Flash thinking ( #2752 )
2025-04-18 12:26:17 -07:00
Nico Bihan
b5a77e34a4
Fixes maximum token limit for Gemini provider 2.5 pro exp ( #2737 )
...
Corrects the maximum token limit for the "gemini-2.5-pro-exp-03-25" model, ensuring accurate configuration.
2025-04-18 00:35:06 -04:00
Nico Bihan
bea36c8977
Gemini 2.5 Flash Preview fix Max Tokens Count ( #2735 )
...
Gemini 2.5 Flash Preview fix Max Tokens
2025-04-17 23:19:40 -04:00
Matt Rubens
d86d601104
Add gemini 2.5 flash preview ( #2720 )
2025-04-17 16:39:00 -04:00
Matt Rubens
43668e0429
Add support for different reasoning effort ( #2692 )
2025-04-16 14:16:51 -04:00
