feat(pricing): add openrouter/z-ai/glm-5.2 and sakana/fugu-ultra to cost map

[blakeaa827]

What

Adds cost-map entries for two OpenRouter models currently missing from the map:

• openrouter/z-ai/glm-5.2 — $0.91 / $2.86 per 1M in/out, cache-read $0.169/1M, reasoning + tools + prompt caching
• openrouter/sakana/fugu-ultra — $5.00 / $30.00 per 1M in/out, cache-read $0.50/1M, reasoning + tools + prompt caching + vision

Why

Without a cost-map entry LiteLLM can't price these models. It's most visible on streaming OpenRouter calls, where the provider-returned cost isn't currently used (see #11626 / #16162) and LiteLLM falls back to the static map — reporting $0.

Values

Prices, context windows, and capability flags come from the OpenRouter model API (GET /api/v1/models): glm-5.2 = 1,048,576 ctx; fugu-ultra = 1,000,000 ctx (multimodal text+image). Entries mirror the existing openrouter/z-ai/glm-5.1 shape. Only model_prices_and_context_window.json is edited, matching how glm-5.1 was added (the packaged backup is synced by maintainers at release time).

Pre-submission checklist

• Data-only change (new cost-map entries); no code paths touched. No unit test, consistent with prior model-pricing additions.
• JSON validated.

🤖 Generated with Claude Code

[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]

Greptile Summary

Adds cost-map entries for two previously unmapped OpenRouter models — openrouter/z-ai/glm-5.2 and openrouter/sakana/fugu-ultra — so LiteLLM can price these models instead of reporting $0 on streaming calls that lack provider-returned cost data.

• openrouter/z-ai/glm-5.2: $0.91/$2.86 per 1M in/out, cache-read $0.169/1M, 1,048,576 token context; structure and field set mirror the existing glm-5.1 entry.
• openrouter/sakana/fugu-ultra: $5.00/$30.00 per 1M in/out, cache-read $0.50/1M, 1,000,000 token context; adds supports_vision: true matching the model's multimodal capability.

Confidence Score: 5/5

Pure data addition with no logic changes; all numeric values are internally consistent and match the PR description.

Both entries follow the established pattern for OpenRouter models (field order, token representation, provider flag), all per-token prices convert correctly to the advertised per-million figures, and only the JSON price file is touched.

No files require special attention.

Important Files Changed

Filename	Overview
model_prices_and_context_window.json	Adds two new OpenRouter model entries (z-ai/glm-5.2 and sakana/fugu-ultra) with correct pricing, context window, and capability flags; structure mirrors the existing glm-5.1 entry and all numeric values are internally consistent with the PR description.

_{Reviews (1): Last reviewed commit: "feat(pricing): add openrouter/z-ai/glm-5..." | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[devin-ai-integration]

Superseded by rolling registry audit #37902, which reverified and absorbed the provider-documented metadata changes at field level