[Feat] add ElevenLabs eleven_v3 and eleven_multilingual_v2 to model cost map

[Chesars]

Relevant issues

N/A - New model registration

Pre-Submission checklist

• My PR passes all unit tests on make test-unit
• My PR's scope is as isolated as possible, it only solves 1 specific problem

Type

🆕 New Feature
📖 Documentation

Changes

Register ElevenLabs TTS models in the model cost map for cost tracking:

• elevenlabs/eleven_v3 — most expressive model, 70+ languages, audio tags support for sound effects and pauses
• elevenlabs/eleven_multilingual_v2 — default TTS model, 29 languages, stable and production-ready

Both models priced at $0.18/1000 characters (Scale plan, 1 credit per character).

No code changes needed — the ElevenLabs integration is already model-agnostic and passes any model_id directly to the API.

Documentation updates

• Added supported models table to ElevenLabs provider docs
• Added example using eleven_v3 with audio tags (<sfx>, <pause>)
• Added tip about model-agnostic support

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
litellm	Ready	Preview, Comment	Feb 5, 2026 9:01pm

[greptile-apps]

Greptile Overview

Greptile Summary

This PR registers two new ElevenLabs TTS models (elevenlabs/eleven_v3 and elevenlabs/eleven_multilingual_v2) in the model cost map for cost tracking purposes.

Key Changes:

• Added elevenlabs/eleven_v3 to cost map with $0.18/1000 characters pricing
• Added elevenlabs/eleven_multilingual_v2 to cost map with same pricing structure
• Updated documentation with supported models table and usage examples
• Added example demonstrating eleven_v3 audio tags feature (<sfx> and <pause>)
• Both JSON files (main and backup) kept in sync correctly

Technical Notes:

• Pricing calculation is correct: $0.18 per 1000 characters = 0.00018 per character
• Model entries follow existing ElevenLabs pattern with proper metadata
• Integration remains model-agnostic (models passed directly as model_id to API)
• No code changes required - existing ElevenLabs integration handles any model

Confidence Score: 5/5

• This PR is safe to merge with no concerns
• Clean registration of new models with correct pricing structure, consistent formatting, properly maintained backup files, and helpful documentation updates with no code changes required
• No files require special attention

Important Files Changed

Filename	Overview
model_prices_and_context_window.json	Added two new ElevenLabs TTS models (eleven_v3 and eleven_multilingual_v2) with correct pricing structure and metadata
litellm/model_prices_and_context_window_backup.json	Identical changes to main JSON file - backup file correctly maintained in sync
docs/my-website/docs/providers/elevenlabs.md	Added supported models table, usage example with audio tags feature, and tip about model-agnostic support

Sequence Diagram

sequenceDiagram
    participant User
    participant LiteLLM
    participant CostMap as Model Cost Map
    participant ElevenLabs as ElevenLabs API

    User->>LiteLLM: speech(model="elevenlabs/eleven_v3", input="text")
    LiteLLM->>CostMap: Lookup model pricing
    CostMap-->>LiteLLM: Return $0.00018 per character
    LiteLLM->>ElevenLabs: POST /v1/text-to-speech/{voice_id}<br/>{"text": "...", "model_id": "eleven_v3"}
    ElevenLabs-->>LiteLLM: Audio stream
    LiteLLM->>CostMap: Calculate cost based on input length
    LiteLLM-->>User: Return audio + cost tracking

Loading

[greptile-apps]

_{3 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[ishaan-jaff]

lgtm