chore(prices): sync OpenRouter prices: 2 models - #42006
Conversation
|
|
 |
|
openrouter/qwen/qwen-plus-2025-07-28 supports_prompt_caching should stay false, not true: https://openrouter.ai/api/v1/models lists no input_cache_read for that exact id |
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
42819cc to
4cd9ff6
Compare
berriai-litellm-provider-info-sync
Bot
changed the title
4cd9ff6 to
b12327b
Compare
openrouter/deepseek/deepseek-v4-flash: input_cost_per_token, output_cost_per_token, cache_read_input_token_cost openrouter/z-ai/glm-5.2: input_cost_per_token, output_cost_per_token, cache_read_input_token_cost
berriai-litellm-provider-info-sync
Bot
changed the title
b12327b to
aa8bbdb
Compare
2 participants
Updates 2 models, each read from its provider's own published prices.
OpenRouter
Updates 2 models from the provider's pricing page (page
75b398df8b9b).Changes
openrouter/deepseek/deepseek-v4-flash: cache_read_input_token_cost $0.008064/1M → $0.00756/1M, input_cost_per_token $0.04032/1M → $0.0378/1M, output_cost_per_token $0.08064/1M → $0.0756/1Mopenrouter/z-ai/glm-5.2: cache_read_input_token_cost $0.10296/1M → $0.12064/1M, input_cost_per_token $0.5544/1M → $0.6496/1M, output_cost_per_token $1.7424/1M → $2.0416/1MNo litellm key or field (19)
Rows the page prices that litellm has no key or field for. Nothing to apply; listed so the omission is deliberate, not missed.
qwen/qwen3.7-flash: override tier at 32000 prompt tokens has no litellm fieldqwen/qwen3.7-flash: override tier at 256000 prompt tokens has no litellm fieldopenrouter/auto-beta: OpenRouter router, not a modelopenrouter/fusion: OpenRouter router, not a modelqwen/qwen3.7-plus: override tier at 256000 prompt tokens has no litellm fieldqwen/qwen3.5-plus-20260420: override tier at 256000 prompt tokens has no litellm fieldqwen/qwen3.6-flash: override tier at 256000 prompt tokens has no litellm fieldopenrouter/pareto-code: OpenRouter router, not a modelqwen/qwen3.6-plus: override tier at 256000 prompt tokens has no litellm fieldqwen/qwen3.5-plus-02-15: override tier at 256000 prompt tokens has no litellm fieldqwen/qwen3-max-thinking: override tier at 32000 prompt tokens has no litellm fieldopenrouter/free: OpenRouter router, not a modelopenrouter/bodybuilder: OpenRouter router, not a modelqwen/qwen3-max: override tier at 32000 prompt tokens has no litellm fieldqwen/qwen3-coder-plus: override tier at 32000 prompt tokens has no litellm fieldqwen/qwen3-coder-flash: override tier at 32000 prompt tokens has no litellm fieldqwen/qwen-plus-2025-07-28: override tier at 256000 prompt tokens has no litellm fieldqwen/qwen-plus: override tier at 256000 prompt tokens has no litellm fieldopenrouter/auto: OpenRouter router, not a model2 rows the source lists without a serverless token price (dedicated or unpriced models) were not read.
Opened by the litellm-providers price sync. Every price is read from the provider's own published source and gated before it is applied; the audit trail for each value is in the portal's sync history.
Note
Low Risk
Data-only updates to published per-token prices in the cost map; billing estimates change for those two models but no runtime logic is modified.
Overview
Syncs OpenRouter token pricing in
model_prices_and_context_window.jsonand its backup for two models, with no capability or limit changes.openrouter/z-ai/glm-5.2input, output, and cache-read costs are increased to match the provider’s current published rates.openrouter/deepseek/deepseek-v4-flashinput, output, and cache-read costs are lowered by the same kind of sync.Only per-token cost fields change; everything else on those entries is unchanged.
Reviewed by Cursor Bugbot for commit aa8bbdb. Bugbot is set up for automated code reviews on this repo. Configure here.