feat(model-prices): add missing OpenAI dated snapshots for deep research, search preview, audio mini and computer use

[devin-ai-integration]

TLDR

Problem this solves:

• Six OpenAI dated snapshots that /v1/models returns had no cost map entry
• Calls pinned to those snapshot names went unpriced

How it solves it:

• Adds each snapshot as a copy of its base model entry
• Sets source to the official OpenAI model page listing the snapshot

User Flow

Before: a developer pins a dated OpenAI snapshot and the request is logged with no cost

1. They send POST https://litellm-domain/v1/responses with "model": "o3-deep-research-2025-06-26"
2. The call succeeds but https://litellm-domain/ui/?page=logs shows $0 spend for it

After: the same request is priced at the snapshot's published rates

1. They send the same POST with "model": "o3-deep-research-2025-06-26"
2. The logs page shows spend at $10 / $40 per 1M input / output tokens

Changes by provider

OpenAI

Each new entry copies its base row (rates, limits, endpoints, flags) with only source changed and deprecation_date left off. Every snapshot is listed under "Snapshots" on its model page, and the page's pricing block matches the base row

New key	Base row	Input / cached / output per 1M	Source
o3-deep-research-2025-06-26	o3-deep-research	$10 / $2.50 / $40	https://developers.openai.com/api/docs/models/o3-deep-research
o4-mini-deep-research-2025-06-26	o4-mini-deep-research	$2 / $0.50 / $8	https://developers.openai.com/api/docs/models/o4-mini-deep-research
gpt-audio-mini-2025-10-06	gpt-audio-mini	$0.60 / n/a / $2.40	https://developers.openai.com/api/docs/models/gpt-audio-mini
gpt-4o-search-preview-2025-03-11	gpt-4o-search-preview	$2.50 / n/a / $10	https://developers.openai.com/api/docs/models/gpt-4o-search-preview
gpt-4o-mini-search-preview-2025-03-11	gpt-4o-mini-search-preview	$0.15 / n/a / $0.60	https://developers.openai.com/api/docs/models/gpt-4o-mini-search-preview
computer-use-preview-2025-03-11	computer-use-preview	$3 / n/a / $12	https://developers.openai.com/api/docs/models/computer-use-preview

Checked and left alone

Anthropic, xAI, DeepSeek, Together, Voyage and Gemini image models matched their official API or pricing pages. Mistral codestral-2508 stays at 128k because https://docs.mistral.ai/models/codestral-25-08 says 128k even though /v1/models reports 256k. Gemini image models keep their limits because the ai.google.dev model pages match the registry

Not added because no per-token price is published: Gemini antigravity-preview-* (billed at the underlying model's rates per https://ai.google.dev/gemini-api/docs/pricing) and aqa, Groq allam-2-7b (missing from https://groq.com/pricing and https://console.groq.com/docs/models), and Fireworks qwen3p8-2p4t-a95b (on-demand only per https://fireworks.ai/models/fireworks/qwen3p8-2p4t-a95b)

Perplexity pplx-decider-v1-27b should be $0.02 per 1M input per https://docs.perplexity.ai/getting-started/pricing, but the open PR #45476 already fixes that row and adds v1.1, so it is not repeated here

Other open registry PRs reviewed

Nothing was absorbed. #36274 (Azure gpt-5.6 terra/luna) conflicts with the Azure Retail Prices API, which shows global terra at $2 / $12 and luna at $0.20 / $1.20, the same as main. #43879 duplicates moonshot/kimi-k2.5 as a bare key with a conflicting 8192 output limit. #40368 edits Bedrock Claude 3 Haiku rows that are no longer on main. #30775 adds Grok 4.3 on Bedrock Converse, but the AWS supported-models page does not list it (main already has bedrock_mantle/xai.grok-4.3). #32117 (Cohere embed light v3) and #45502 (SCX.ai) list prices that the current official pricing pages do not show. Code or behavior PRs that only touch the JSON incidentally were left alone, including #45476, #43390, #33922, #43363, #41857, #29898, #44651, #44167, #43457 and #40755

Pre-Submission checklist

• I have added meaningful tests
• The handful of test files covering my change pass locally, e.g. uv run pytest tests/unit/<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

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

LITELLM_LOCAL_MODEL_COST_MAP=True uv run --no-sync pytest -q tests/unit/test_model_prices_schema.py tests/unit/test_model_cost_aliases.py returns 92 passed. python3 ci_cd/check_files_match.py passes, and uv run --no-sync python ci_cd/cost_map_guard.py --base origin/main passes

Link to Devin session: https://app.devin.ai/sessions/5786090f2b044d3789a4f41c4dc5996f
Open in Devin Desktop: https://app.devin.ai/desktop/session/5786090f2b044d3789a4f41c4dc5996f?variant=devin

[devin-ai-integration]

I'll fix CI failures and address comments from users with write access. I'll skip comments containing "(aside)".

• 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.}

[greptile-apps]

Confidence Score: 5/5

[Medium impact] This PR appears safe to merge.

Summary

Adds six dated OpenAI snapshots to both model price maps.

• Dated OpenAI snapshots now have matching price entries.

Acknowledgment: devin-ai-integration[bot] explicitly states that deprecation_date is intentionally omitted.

_{Reviews (1) · Last reviewed commit: "feat(model-prices): add missing OpenAI d..." · Reviewed by Greptile}

[codspeed]

Merging this PR will not alter performance

✅ 31 untouched benchmarks

---

_{Comparing litellm_model_prices_openai_snapshots_20261010 (0e9eb69) with main (7831d8a)}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!