Skip to content

feat(model_prices): add deepseek-v3 to model price maps (#43660) - #43753

Open
VimalN2005 wants to merge 1 commit into
BerriAI:mainfrom
VimalN2005:feat/add-deepseek-v3-model-prices
Open

VimalN2005 wants to merge 1 commit into
BerriAI:mainfrom
VimalN2005:feat/add-deepseek-v3-model-prices

Conversation

@VimalN2005

Copy link
Copy Markdown

Overview

Closes #43660.

This PR adds the missing bare deepseek-v3 model entry to both model_prices_and_context_window.json and litellm/model_prices_and_context_window_backup.json to maintain parity with other DeepSeek models (such as deepseek-chat, deepseek-reasoner, and deepseek-v4-flash), which possess both bare and provider-prefixed entries.

Pricing Details

Sourced from DeepSeek Official Pricing:

  • Input token cost (cache miss): .27 / 1M tokens (2.7e-07)
  • Input token cost (cache hit): .07 / 1M tokens (7e-08)
  • Output token cost: .10 / 1M tokens (1.1e-06)
  • Max input tokens: 65,536
  • Max output tokens: 8,192
  • Provider: deepseek
  • Mode: chat
  • Capabilities: function calling, tool choice, assistant prefill, prompt caching

Changes

  • model_prices_and_context_window.json: Added deepseek-v3 entry.
  • litellm/model_prices_and_context_window_backup.json: Added deepseek-v3 entry.
  • ests/test_litellm/test_deepseek_v3_costs.py: Added regression test suite verifying cost map entries, get_model_info('deepseek-v3'), and completion_cost calculations.

Verification

  • python -m pytest tests/test_litellm/test_model_prices_schema.py -v: 38 passed
  • python -m pytest tests/test_litellm/test_deepseek_v3_costs.py -v: 4 passed

uff check tests/test_litellm/test_deepseek_v3_costs.py: Clean (0 errors)

@greptile-apps

greptile-apps Bot commented Sep 29, 2026 •

Copy link
Copy Markdown
Contributor

RetriggerConfidence Score: 4/5

[Medium risk] Adds pricing data for a new model to the price maps.

The PR should not merge until unprefixed DeepSeek V3 requests work and the explicit test requirement is satisfied.

Findings

  1. P1 Bare model cannot route ▶
  2. P2 Tests pin vendor prices ▶
  3. P2 Cached pricing remains untested ▶

Summary

Adds an unprefixed DeepSeek V3 entry to the main and backup price maps, with metadata and cost tests.

  • Bare model requests still lack provider inference
  • Tests pin vendor prices and omit cached-usage pricing

Reviews (1) · Last reviewed commit: "feat(model_prices): add deepseek-v3 to m..."

Comment thread model_prices_and_context_window.json
Comment thread tests/test_litellm/test_deepseek_v3_costs.py
Comment thread tests/test_litellm/test_deepseek_v3_costs.py
@codecov

codecov Bot commented Sep 29, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

@codspeed

codspeed Bot commented Sep 29, 2026

Copy link
Copy Markdown
Contributor

Merging this PR will not alter performance

✅ 31 untouched benchmarks


Comparing VimalN2005:feat/add-deepseek-v3-model-prices (aa5d6eb) with main (f4a7c04)

Open in CodSpeed

@VimalN2005

Copy link
Copy Markdown
Author

Added deepseek-v3 pricing to both maps, covered schema tests, and verified cost calculations. Ready for review when you have time.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Add "deepseek-v3" in "model_prices_and_context_window.json"

1 participant