Automated check of the backend's hardcoded model lists against the live model API (https://ai-models.puter.work/<provider>).
A backend model is reported as deprecated only when neither its id nor any of its aliases appears anywhere in the upstream name set (upstream id + name + aliases), case-insensitive. Pairs with zero deprecations are omitted.
Extraction sanity check: every backend file's record count matched its puterId: count (or factory-call count for zai), so no pair failed extraction.
chat/claude (1)
Notes: upstream Anthropic list is 10 models (claude-opus-5, claude-sonnet-5, claude-fable-5, claude-opus-4-8, claude-opus-4-7, claude-sonnet-4-6, claude-opus-4-6, claude-opus-4-5-20251101, claude-haiku-4-5-20251001, claude-sonnet-4-5-20250929). Opus 4.1 is the only 4.x entry no longer present.
chat/openai (2)
Notes: o3 and o3-mini are still served; only the -pro variant is gone. No gpt-4.5* entry remains upstream.
chat/mistral (2)
magistral-medium-2509open-mistral-nemo-2407
Notes: upstream still serves magistral-small-latest (as an alias of mistral-small-2603), but there is no magistral-medium under any id or alias. No Nemo model remains upstream.
chat/xai (10)
grok-3grok-3-fastgrok-3-minigrok-3-mini-fastgrok-4-1-fastgrok-4-1-fast-non-reasoninggrok-4-fastgrok-4-fast-non-reasoninggrok-4-0709grok-4
Notes: this is a large drop, so it was sanity-checked against aliases. Upstream xAI now lists only 10 models: grok-4.20-0309-non-reasoning, grok-4.20-0309-reasoning, grok-4.20-multi-agent-0309, grok-4.3, grok-4.5, grok-build-0.1, grok-imagine-image, grok-imagine-image-quality, grok-imagine-video, grok-imagine-video-1.5. The current flagships (grok-4.5, grok-4.3) are present, so the endpoint is not returning a broken set. grok-code-fast-1 is not deprecated — it is an alias of grok-build-0.1. The entire Grok 3 line and every Grok 4 / 4.1 variant is genuinely no longer offered.
chat/alibaba (2)
qwen3-coder-30b-a3b-instructqwen3-vl-30b-a3b
Notes: upstream still serves qwen3-coder-plus, qwen3-coder-flash, qwen3-coder-480b-a35b-instruct, qwen3-coder-next, qwen3-vl-plus, and qwen3-vl-flash — just not these two 30B-A3B variants.
tts/xai (1)
Notes: low confidence — the xAI endpoint lists no TTS models at all (chat + imagine image/video only), so this may be a Puter-internal id for xAI's voice API rather than an upstream-listed model. Worth confirming by hand before acting on it.
video/gemini (2)
veo-3.1-fast-generate-previewveo-3.1-lite-generate-preview
Notes: veo-3.1-generate-preview is listed upstream, which confirms Veo models do appear in the Gemini listing — so the absence of the -fast and -lite variants is a real signal rather than an endpoint that omits video models.
Verify manually — chat/zai: the zai endpoint returns only base GLM models (glm-4.5, glm-4.5-air, glm-4.6, glm-4.7, glm-5, glm-5-turbo, glm-5.1, glm-5.2) and lists no -flash/-flashx/-x/-airx/vision variants at all. That makes the following 12 backend models look deprecated, but they most likely are not — please confirm against Z.ai's own docs rather than this endpoint: glm-4.7-flashx, glm-4.7-flash, glm-4.5-x, glm-4.5-airx, glm-4.5-flash, glm-4-32b-0414-128k, glm-5v-turbo, glm-4.6v, glm-4.6v-flashx, glm-4.6v-flash, glm-4.5v, autoglm-phone-multilingual.
No deprecations found for: chat/gemini, chat/deepseek, chat/moonshot, chat/minimax, image/gemini, image/openai, image/xai, tts/gemini, tts/openai, video/openai.
This check was read-only; no files were modified.
Automated check of the backend's hardcoded model lists against the live model API (
https://ai-models.puter.work/<provider>).A backend model is reported as deprecated only when neither its
idnor any of itsaliasesappears anywhere in the upstream name set (upstreamid+name+aliases), case-insensitive. Pairs with zero deprecations are omitted.Extraction sanity check: every backend file's record count matched its
puterId:count (or factory-call count for zai), so no pair failed extraction.chat/claude (1)
claude-opus-4-1-20250805Notes: upstream Anthropic list is 10 models (
claude-opus-5,claude-sonnet-5,claude-fable-5,claude-opus-4-8,claude-opus-4-7,claude-sonnet-4-6,claude-opus-4-6,claude-opus-4-5-20251101,claude-haiku-4-5-20251001,claude-sonnet-4-5-20250929). Opus 4.1 is the only 4.x entry no longer present.chat/openai (2)
o3-progpt-4.5-previewNotes:
o3ando3-miniare still served; only the-provariant is gone. Nogpt-4.5*entry remains upstream.chat/mistral (2)
magistral-medium-2509open-mistral-nemo-2407Notes: upstream still serves
magistral-small-latest(as an alias ofmistral-small-2603), but there is nomagistral-mediumunder any id or alias. No Nemo model remains upstream.chat/xai (10)
grok-3grok-3-fastgrok-3-minigrok-3-mini-fastgrok-4-1-fastgrok-4-1-fast-non-reasoninggrok-4-fastgrok-4-fast-non-reasoninggrok-4-0709grok-4Notes: this is a large drop, so it was sanity-checked against aliases. Upstream xAI now lists only 10 models:
grok-4.20-0309-non-reasoning,grok-4.20-0309-reasoning,grok-4.20-multi-agent-0309,grok-4.3,grok-4.5,grok-build-0.1,grok-imagine-image,grok-imagine-image-quality,grok-imagine-video,grok-imagine-video-1.5. The current flagships (grok-4.5,grok-4.3) are present, so the endpoint is not returning a broken set.grok-code-fast-1is not deprecated — it is an alias ofgrok-build-0.1. The entire Grok 3 line and every Grok 4 / 4.1 variant is genuinely no longer offered.chat/alibaba (2)
qwen3-coder-30b-a3b-instructqwen3-vl-30b-a3bNotes: upstream still serves
qwen3-coder-plus,qwen3-coder-flash,qwen3-coder-480b-a35b-instruct,qwen3-coder-next,qwen3-vl-plus, andqwen3-vl-flash— just not these two 30B-A3B variants.tts/xai (1)
xai-ttsNotes: low confidence — the xAI endpoint lists no TTS models at all (chat + imagine image/video only), so this may be a Puter-internal id for xAI's voice API rather than an upstream-listed model. Worth confirming by hand before acting on it.
video/gemini (2)
veo-3.1-fast-generate-previewveo-3.1-lite-generate-previewNotes:
veo-3.1-generate-previewis listed upstream, which confirms Veo models do appear in the Gemini listing — so the absence of the-fastand-litevariants is a real signal rather than an endpoint that omits video models.Verify manually — chat/zai: the zai endpoint returns only base GLM models (
glm-4.5,glm-4.5-air,glm-4.6,glm-4.7,glm-5,glm-5-turbo,glm-5.1,glm-5.2) and lists no-flash/-flashx/-x/-airx/vision variants at all. That makes the following 12 backend models look deprecated, but they most likely are not — please confirm against Z.ai's own docs rather than this endpoint:glm-4.7-flashx,glm-4.7-flash,glm-4.5-x,glm-4.5-airx,glm-4.5-flash,glm-4-32b-0414-128k,glm-5v-turbo,glm-4.6v,glm-4.6v-flashx,glm-4.6v-flash,glm-4.5v,autoglm-phone-multilingual.No deprecations found for: chat/gemini, chat/deepseek, chat/moonshot, chat/minimax, image/gemini, image/openai, image/xai, tts/gemini, tts/openai, video/openai.
This check was read-only; no files were modified.