[Perplexity AI] update models #15667
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎ |
WalkthroughThis pull request updates version numbers and a constants array across three components. The chat-completions module’s exported version is bumped from "0.0.4" to "0.0.5", and the package version is updated from "0.1.3" to "0.1.4". Additionally, the Changes
Possibly related PRs
Suggested labels
Suggested reviewers
Poem
Warning There were issues while running some tools. Please review the errors and either fix the tool’s configuration or disable the tool if it’s a critical failure. 🔧 ESLint
components/perplexity/actions/chat-completions/chat-completions.mjsOops! Something went wrong! :( ESLint: 8.57.1 Error [ERR_MODULE_NOT_FOUND]: Cannot find package 'jsonc-eslint-parser' imported from /eslint.config.mjs components/perplexity/common/constants.mjsOops! Something went wrong! :( ESLint: 8.57.1 Error [ERR_MODULE_NOT_FOUND]: Cannot find package 'jsonc-eslint-parser' imported from /eslint.config.mjs ✨ Finishing Touches
Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media? 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🔭 Outside diff range comments (1)
components/perplexity/common/constants.mjs (1)
1-13:⚠️ Potential issue❓ Verification inconclusive
⚠️ Breaking Change: Model names have been updated.The removal of specific model versions (
llama-3.1-sonar-small-128k-online,llama-3.1-sonar-large-128k-online) in favor of generalized identifiers could break existing implementations. Please ensure:
- This change is documented in the changelog
- Migration instructions are provided for users
- The version bump (0.1.3 → 0.1.4) adequately reflects this breaking change - consider a major version bump (1.0.0) instead
Let's check for any existing usages that might be affected:
🏁 Script executed:
#!/bin/bash # Search for any usage of the old model names that might need updating rg -i "llama-3\.1-sonar-(small|large)-128k-online" --type js --type mjsLength of output: 101
Breaking Change: Verify Deprecated Model Name Usages
The update removes specific model identifiers (
llama-3.1-sonar-small-128k-onlineandllama-3.1-sonar-large-128k-online) in favor of generalized names. This could break existing implementations that might still reference these explicit names.Action Items:
- Documentation: Ensure the changelog clearly documents this breaking change and provide detailed migration instructions for users.
- Versioning: Reassess whether the minor version bump (0.1.3 → 0.1.4) sufficiently communicates the breaking nature of this change, or if a major version bump (e.g., 1.0.0) is more appropriate.
- Verification: Confirm that there are no residual hard-coded usages of the old model names in the codebase. The original verification command didn’t work for
.mjsfiles. Please re-run the following updated command to search in both.jsand.mjsfiles:#!/bin/bash rg -i "llama-3\.1-sonar-(small|large)-128k-online" -g "*.{js,mjs}"Review any findings, and if old references are still present, update them accordingly.
🧹 Nitpick comments (1)
components/perplexity/common/constants.mjs (1)
1-13: Consider documenting the model capabilities and differences.The new model naming scheme (
sonarvssonar-provssonar-reasoningvssonar-reasoning-pro) suggests different capabilities or performance characteristics. Consider adding JSDoc comments to document:
- The differences between base and pro models
- The specific capabilities of reasoning models
- Any migration guidance for users of the old model names
export default { + /** + * Available Perplexity AI models: + * - sonar: Base model for general use + * - sonar-pro: Enhanced version with improved capabilities + * - sonar-reasoning: Specialized for reasoning tasks + * - sonar-reasoning-pro: Enhanced reasoning model + * @see https://docs.perplexity.ai/docs/model-cards + */ MODELS: [ "sonar", "sonar-pro", "sonar-reasoning", "sonar-reasoning-pro", ],
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
pnpm-lock.yamlis excluded by!**/pnpm-lock.yaml
📒 Files selected for processing (3)
components/perplexity/actions/chat-completions/chat-completions.mjs(1 hunks)components/perplexity/common/constants.mjs(1 hunks)components/perplexity/package.json(1 hunks)
✅ Files skipped from review due to trivial changes (2)
- components/perplexity/package.json
- components/perplexity/actions/chat-completions/chat-completions.mjs
⏰ Context from checks skipped due to timeout of 90000ms (3)
- GitHub Check: pnpm publish
- GitHub Check: Verify TypeScript components
- GitHub Check: Publish TypeScript components
🔇 Additional comments (1)
components/perplexity/common/constants.mjs (1)
3-6:✅ Verification successful
Verify the updated model identifiers against Perplexity's API documentation.
The change from specific model versions to generalized identifiers could impact existing integrations. Let's verify these new model names are correct.
🌐 Web query:
What are the available model names in Perplexity AI's API?💡 Result:
Perplexity AI's API provides access to multiple advanced language models, including both proprietary and open-source options. Here's a breakdown of available models based on current documentation:
Core Proprietary Models (Sonar Series)
Available until February 22, 2025 (deprecation notice)[7]:
llama-3.1-sonar-small-128k-onlinellama-3.1-sonar-large-128k-onlinellama-3.1-sonar-huge-128k-onlineCurrent recommended replacements[7][10]:
sonar(127k context)sonar-pro(200k context)sonar-reasoning(127k context)sonar-reasoning-pro(127k context)Open-Source Foundation Models
Available through pplx-api and third-party integrations[4][9]:
mistral-7b-instructcodellama-34b-instructllama-2-13b-chatllama-2-70b-chatopenhermes-2-mistral-7bopenhermes-2.5-mistral-7bSpecialized Variants
Additional optimized versions available via API[9][10]:
pplx-7b-chat pplx-70b-chat pplx-7b-online pplx-70b-online pplx-7b-chat-alpha pplx-70b-chat-alphaUsage Notes
- Model access requires an API key through Perplexity's authentication system[2][4]
- Default model is
sonarwhen not specified[3]- Full list updates dynamically - check
/modelsendpoint for latest availability[5]- Citations can be enabled via
return_citations=Trueparameter (beta feature)[9]Developers can access these models through:
client.chat.completions.create( model="perplexity/sonar-pro", messages=[...] )For complete specifications and parameters, consult Perplexity's official model cards documentation[10].
Citations:
- 1: https://www.perplexity.ai/hub/technical-faq/what-advanced-ai-models-does-perplexity-pro-unlock
- 2: https://docs.superblocks.com/integrations/integrations-library/perplexity
- 3: https://spiceai.org/docs/components/models/perplexity
- 4: https://www.perplexity.ai/hub/blog/introducing-pplx-api
- 5: https://apipie.ai/docs/Models/Perplexity
- 6: https://www.perplexity.ai/hub/faq/pplx-api
- 7: https://docs.perplexity.ai/changelog/changelog
- 8: https://www.postman.com/ai-engineer/generative-ai-large-language-model-apis/documentation/lrm4umz/perplexity-ai-api
- 9: https://docs.litellm.ai/docs/providers/perplexity
- 10: https://docs.perplexity.ai/guides/model-cards
Verified model identifiers update – no issues detected
The new model identifiers
"sonar","sonar-pro","sonar-reasoning", and"sonar-reasoning-pro"now align with Perplexity AI’s current recommendations per their documentation. This update replaces the older, version-specific models with more generalized identifiers. Please ensure that integration tests and documentation are updated accordingly, as this change could affect clients relying on previous model names.
WHY
Summary by CodeRabbit
Chores
Refactor