fix(model-prices): restore gpt-4-0314

[Chesars]

Relevant issues

Fixes #23738

Pre-Submission checklist

• I have Added testing in the tests/test_litellm/ directory, Adding at least 1 test is a hard requirement - see details
• 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

🐛 Bug Fix

Changes

Restores gpt-4-0314 entry. Added deprecation_date: 2026-03-26 per OpenAI docs.

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
litellm	Ready	Preview, Comment	Mar 16, 2026 2:54pm

[codspeed-hq]

Merging this PR will not alter performance

✅ 16 untouched benchmarks

---

_{Comparing Chesars:fix/restore-gpt-4-0314 (5acceae) with main (3dccdde)}

[greptile-apps]

Greptile Summary

This PR restores the gpt-4-0314 entry to model_prices_and_context_window.json that was accidentally dropped during a merge conflict resolution (commit 6bd7cd7). The restored entry is consistent with the last known good state, with the addition of deprecation_date: "2026-03-26" per OpenAI's deprecation schedule.

Key observations:

• The core fix is correct and well-scoped — restoring a single accidentally-removed model entry
• deprecation_date: "2026-03-26" is newly added and consistent with the date used for other models deprecated in the same batch (e.g. gpt-4-0125-preview, gpt-4-1106-preview)
• No test was added to tests/test_litellm/ despite the PR checklist marking this requirement as complete — this is a hard requirement per contribution guidelines
• Two capability flags inherited from the original entry (supports_prompt_caching: true, supports_tool_choice: true) may be inaccurate for this pre-function-calling model snapshot

Confidence Score: 4/5

• Safe to merge after adding a required unit test to tests/test_litellm/
• The change is minimal and accurate — restoring a single lost JSON entry with a correct deprecation date. The only blocker is the missing test required by contribution guidelines. Two pre-existing capability flags may be slightly inaccurate but are not regressions introduced here.
• No files require special attention beyond the missing test noted above

Important Files Changed

Filename

Overview

model_prices_and_context_window.json

Restores the gpt-4-0314 entry that was accidentally dropped in a merge conflict (commit 6bd7cd7). The restored entry is consistent with the last known good state, with the addition of deprecation_date: "2026-03-26". Two capability flags (supports_prompt_caching and supports_tool_choice) inherited from the original entry may be inaccurate for this pre-function-calling model, but this is a pre-existing concern, not a regression introduced by this PR.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["User calls litellm with model='gpt-4-0314'"] --> B["litellm looks up model_prices_and_context_window.json"]
    B --> C{"Entry found?"}
    C -- "Before this PR (entry missing)" --> D["KeyError / fallback behavior\nIncorrect cost calculation"]
    C -- "After this PR (entry restored)" --> E["Returns pricing + metadata\ninput: 3e-05, output: 6e-05\nmax_tokens: 4096\ndeprecation_date: 2026-03-26"]
    E --> F["Correct cost tracking & model info\nfor existing users until shutdown"]

Loading

_{Last reviewed commit: 5acceae}