Repository navigation
chore: sync capabilities from the live catalog (2026-10-20): GPT-6.1 Sol, Mistral Large 4 - #7
Merged
Merged
Conversation
…Sol, Mistral Large 4 Regenerate model-capabilities.ts with scripts/sync-model-capabilities.mjs from GET https://blockrun.ai/api/v1/models (115 ids, 88 text models). Added: openai/gpt-6.1-sol and mistral/mistral-large-4 (both vision-capable per the catalog). Neither has tool-probe evidence, so both are written supportsTools: false. They can serve plain chat but not tool-required turns until a probe admits them. No other capability entries changed. config.ts tier chains were checked against the live list. Every id they name is still served, so no chain is retargeted. The free rungs (nemotron-3.5-lightning and nemotron-3-nano-omni) currently report available: false on Base, but that is the gateway's free-tier circuit breaker, not a delisting. Hosts handle it with RouterOptions.unavailableModels. model-profiles.generated.json and scripts/probe-tools.json still hold retired ids (mistral-nemotron, nemotron-nano-9b/12b, step-3.7-flash, hy3, gpt-5.3). They are left alone: only the paid probe writes those files, and earlier retirements did not prune them by hand. README snapshot count updated; dist/ rebuilt.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What changed
model-capabilities.tsregenerated withscripts/sync-model-capabilities.mjsfrom the live catalog (88 text models).openai/gpt-6.1-sol(1.05M ctx, 128K out, vision) andmistral/mistral-large-4(524K ctx, 262K out, vision).supportsTools: false. They can serve plain chat but not tool-required turns until probed. The paid probe was not run.dist/rebuilt.Not changed (and why)
config.tstiers: every id in the tier and fallback chains is still on the live list, so nothing routes to a dead model. The free rungsnvidia/nemotron-3.5-lightningandnvidia/nemotron-3-nano-omni-30b-a3b-reasoningcurrently reportavailable: falseon Base. That flag comes from the gateway's free-tier circuit breaker in/api/v1/models, not from a delisting, and hosts already handle it withRouterOptions.unavailableModels.model-profiles.generated.json/scripts/probe-tools.json: these still hold retired ids (nvidia/mistral-nemotron,nvidia/nemotron-nano-9b-v2,nvidia/nemotron-nano-12b-v2-vl,nvidia/step-3.7-flash,tencent/hy3, plusopenai/gpt-5.3in probe-tools). Only the paid probe writes these files, and earlier retirements (for example the hy3 retirement in 4cca4e0) did not prune them by hand. They are inert priors for unlisted ids.Source of truth
GET https://blockrun.ai/api/v1/models, fetched 2026-10-20T04:12Z.Checks
npm run check: typecheck passes, vitest 146/146 (decision snapshot unchanged), tsup build passes.No version bump, tag or publish.