feat(models): add quantization to nebius mappings

[AmineAce]

Adds quantization to Nebius provider mappings across the catalog, per #2946.
Checked every Nebius mapping against Nebius Token Factory's model catalog. Added the field only where the endpoint details panel explicitly showed a "Quantization" value.
Confirmed:

Qwen3-235B-A22B-Instruct-2507 → fp8
Qwen3-32B → fp8
Qwen2.5-VL-72B-Instruct → fp8
Qwen3-30B-A3B-Instruct-2507 → fp8
Qwen3-Next-80B-A3B-Thinking → fp8
Qwen3.5-397B-A17B → fp4
Llama-3_1-Nemotron-Ultra-253B-v1 → fp8
Llama-3.3-70B-Instruct → fp8
gpt-oss-120b → fp4
MiniMax-M2.5 → fp4
gemma-3-27b-it → fp8

Checked but not currently listed on Nebius (20, left unchanged)

QwQ-32B
Qwen3-235B-A22B-Thinking-2507
Qwen3-14B
Qwen3-30B-A3B
Qwen2.5-Coder-7B-fast
Qwen2.5-32B-Instruct
Qwen2.5-72B-Instruct
Qwen2-VL-72B-Instruct
Qwen3-Coder-480B-A35B-Instruct
Qwen3-Coder-30B-A3B-Instruct
Qwen3-30B-A3B-Thinking-2507
DeepSeek-V3
DeepSeek-R1-0528
DeepSeek-V3.2
Meta-Llama-3.1-8B-Instruct
Meta-Llama-3.1-405B-Instruct
Kimi-K2-Instruct
Kimi-K2.5
GLM-5
Hermes-3-Llama-405B

Ran pnpm format; only these 5 files changed.

Summary by CodeRabbit

• Model Updates
  • Added explicit quantization metadata for supported model variants across Alibaba, Google, Meta, MiniMax, and OpenAI providers.
  • Identified FP8 and FP4 configurations for improved model capability and compatibility reporting.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 7ea4e874-450e-4b70-88fa-1a0a6887ff7e

📥 Commits

Reviewing files that changed from the base of the PR and between 22d1af7 and 16a9e45.

📒 Files selected for processing (5)

• packages/models/src/models/alibaba.ts
• packages/models/src/models/google.ts
• packages/models/src/models/meta.ts
• packages/models/src/models/minimax.ts
• packages/models/src/models/openai.ts

---

Walkthrough

Selected Nebius provider mappings now explicitly declare fp8 or fp4 quantization across Alibaba, Google, Meta, Minimax, and OpenAI model definitions.

Changes

Provider quantization metadata

Layer / File(s)	Summary
FP8 provider metadata packages/models/src/models/alibaba.ts, packages/models/src/models/google.ts, packages/models/src/models/meta.ts	Selected Nebius Qwen, Gemma, and Llama provider entries now include quantization: "fp8".
FP4 provider metadata packages/models/src/models/alibaba.ts, packages/models/src/models/minimax.ts, packages/models/src/models/openai.ts	Nebius provider entries for Qwen35, MiniMax, and GPT-OSS now include quantization: "fp4".

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• theopenco/llmgateway#2946: Adds quantization metadata to provider mappings and surfaces it in the UI.
• theopenco/llmgateway#2962: Adds quantization fields to additional provider model mappings.
• theopenco/llmgateway#2963: Adds quantization metadata to Minimax provider mappings.

Suggested reviewers: steebchen, smakosh

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: adding quantization metadata to Nebius model mappings.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}