feat(model_prices): add databricks/databricks-glm-5-3-flash

[devin-ai-integration]

TLDR

Problem this solves:

• Databricks now serves GLM 5.3 Flash (databricks-glm-5-3-flash), missing from the model registry
• Requests to it fail model validation and show no context window info

How it solves it:

• Adds databricks/databricks-glm-5-3-flash to the cost map and backup copy
• 1,048,576 input context, text+image input, function calling, reasoning, prompt caching

Databricks has not published pay-per-token DBU rates for this model yet (it is absent from the foundation-model-serving pricing page as of 2026-08-27, verified live in a browser), so cost fields are intentionally omitted; a follow-up can add them once the rates appear. Capabilities and context length come from the Databricks supported models docs

User Flow

Before: a proxy admin cannot onboard the new Databricks GLM 5.3 Flash endpoint cleanly

1. They add databricks/databricks-glm-5-3-flash to their config and restart the proxy
2. They GET http://localhost:4000/model/info and see no max_input_tokens, no modality or capability data for the model

After: the model resolves from the registry with its real limits and capabilities

1. They add databricks/databricks-glm-5-3-flash to their config and restart the proxy
2. They GET http://localhost:4000/model/info and see max_input_tokens 1048576, supports_vision true, supports_function_calling true, supports_reasoning true

Relevant issues

Model add requested by the team in Slack

Linear ticket

Pre-Submission checklist

Please complete all items before asking a LiteLLM maintainer to review your PR

• I have added meaningful tests (registry-only change, validated against a live proxy below)
• The handful of test files covering my change pass locally, e.g. uv run pytest tests/test_litellm/<your_test_file>.py -v
• My PR passes all required CI/CD checks (e.g., lint, schema.d.ts sync check, etc.)
• My PR's scope is as isolated as possible; it only solves 1 specific problem
• I have received a Greptile Confidence Score of at least 4/5 before requesting a maintainer review (Greptile reviews automatically once the PR is opened; only comment @greptileai to re-request a review after pushing changes)

Delays in PR merge?

If you're seeing a delay in your PR being merged, ping the LiteLLM Team on Slack (#pr-review).

Screenshots / Proof of Fix

After (7ec946c)

Model resolves on a live proxy

Proxy started with the model configured, then:

$ curl -s http://localhost:4010/model/info -H "Authorization: Bearer sk-1234" | python3 -c "..."
{
  "key": "databricks/databricks-glm-5-3-flash",
  "max_input_tokens": 1048576,
  "max_output_tokens": 131072,
  "supports_vision": true,
  "supports_function_calling": true,
  "supports_reasoning": true,
  "supports_prompt_caching": true,
  "litellm_provider": "databricks",
  "mode": "chat"
}

Before this change the same lookup raises: model not found in model_prices_and_context_window.json

Link to Devin session: https://app.devin.ai/sessions/b7c5f03e7a9945bca7944fef42d291bf
Open in Devin Desktop: https://app.devin.ai/desktop/session/b7c5f03e7a9945bca7944fef42d291bf?variant=devin

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[devin-ai-integration]

Out of scope — #38583 has no enterprise label (label list is empty), so no risk label, product-area label, or Linear routing was applied. Author and repo checks did pass.

[greptile-apps]

Greptile Summary

Adds the Databricks GLM 5.3 Flash model to both model metadata registries so it resolves with documented context limits and capabilities.

• Registers a 1,048,576-token input limit and 131,072-token output limit.
• Declares text/image input, text output, function calling, tool choice, reasoning, vision, and prompt caching support.
• Intentionally omits pricing fields until Databricks publishes rates.

Confidence Score: 5/5

The PR appears safe to merge, with both model registries synchronized and no actionable defects identified.

The added metadata conforms to the catalog schema, follows the existing Databricks model-entry structure, and intentionally excludes unavailable pricing data.

Important Files Changed

Filename	Overview
model_prices_and_context_window.json	Adds schema-conforming metadata for the new Databricks model without an identified correctness issue.
litellm/model_prices_and_context_window_backup.json	Keeps the packaged backup registry synchronized with the primary model catalog.

_{Reviews (1): Last reviewed commit: "feat(model_prices): add databricks/datab..." | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[codspeed-hq]

Merging this PR will not alter performance

✅ 31 untouched benchmarks

---

_{Comparing devin_ai_1787870894_databricks_glm_5_3_flash (7ec946c) with litellm_internal_staging (67c7b97)¹}

Footnotes

1. No successful run was found on litellm_internal_staging (44d8436) during the generation of this report, so 67c7b97 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[devin-ai-integration]

Superseded by the rolling registry PR #38560, which absorbs this entry field-for-field (re-verified against the Databricks supported-models docs). Closing to keep a single open registry PR.