Remove the delisted hermes-4-70b; charge OpenRouter's real rate for hy3 - #167
Merged
Merged
Conversation
OpenRouter has delisted nousresearch/hermes-4-70b — it is absent from /api/v1/models, so every request returned "404 No endpoints found". Its pricing page still quotes $0.13/$0.40, so a price check does not establish that a model is served. Removed the enum entry and both lookup aliases; the gateway now rejects the name instead of forwarding a request that cannot succeed. hermes-4-405b is unaffected. hy3 was underpriced: OpenRouter charges $0.132/$0.528 per MTok for tencent/hy3, against 0.0825/0.33 here, so the gateway absorbed 1.6x. Two provider-probe cases in the integration test named models nobody should send any more — the OpenRouter probe used hermes-4-70b (now hy3), the xAI probe used grok-4-fast, retired on 2026-05-15 (now grok-4.3). 457 passed, 6 skipped; make lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QmnaiLePKAQEkNfR8VPfja
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to #166 (merged). Two OpenRouter problems, both found by reading
openrouter.ai/api/v1/modelsrather than pricing pages.hermes-4-70bis gonenousresearch/hermes-4-70bis absent from OpenRouter's catalog entirely — 446 models, no entry. Every request returned:Its pricing page still resolves and still quotes $0.13/$0.40. A listed price does not establish that a model is served.
Removed the enum entry and both lookup aliases, so the gateway rejects the name rather than forwarding a request that cannot succeed.
hermes-4-405bis still served and unaffected.hy3was underpricedOpenRouter charges $0.132/$0.528 per MTok for
tencent/hy3; the registry had0.0825/0.33, so the gateway absorbed 1.6x. Same class as the retired xAI slugs in #166, but from a quiet list-price change rather than a retirement.This matters more than the ratio suggests: hy3 is becoming the chat app's thread-title model, so it is about to be the highest-volume model in the registry.
Also
Two provider-probe cases in the integration test named models nobody should be sending: the OpenRouter probe used
hermes-4-70b(nowhy3), the xAI probe usedgrok-4-fast, one of the slugs retired 2026-05-15 (nowgrok-4.3).test_hermes_4_70b_resolvesbecomestest_hermes_4_70b_is_gone, asserting both spellings raise.Testing
457 passed, 6 skipped;
make lintclean.Related
PCR note
Touches
model_registry.py, so PCRs change.measurements.txtnot updated, matching prior model-registry PRs.🤖 Generated with Claude Code
https://claude.ai/code/session_01QmnaiLePKAQEkNfR8VPfja
Generated by Claude Code