Skip to content
Navigation

Plan event · 15 Nov 2026

All plan events · 32 providers tracked

Charm Hypermodels

Models→NeutralProvider profile →

What moved

19 stored models on every Shred tier — Kimi K3, DeepSeek V4.1 Flash, DeepSeek V4 Pro, GLM-5.3, Qwen3.8-Max, Qwen 3.8 2.4T, DeepSeek V4 Flash, Kimi K2.7 Code, GLM-5.3-Flash, GLM-5.2, MiniMax M3, MiniMax M2.7, Qwen3.8 Flash, Qwen3.8-27B, Qwen3.7-Max, Qwen3.7-Plus, Qwen3.7-Flash, GPT-OSS 120B and Inkling — with "+4 more" over /v1/models → 14 stored models on every Shred tier — Kimi K3, DeepSeek V4.1 Flash, DeepSeek V4 Pro, GLM-5.3, Qwen3.8-Max, Qwen 3.8 2.4T, GLM-5.3-Flash, MiniMax M3, Qwen3.8 Flash, Qwen3.8-27B, GPT-OSS 120B, Inkling, plus Kimi K3.1 Pro ($2.20/$11.00/$0.22 cached read) and Muse Spark Open ($1.40/$4.65/$0.17), each at its own Hyper row — and the array is now the full served-and-tracked roster, so modelsMore is gone

Charm Hyper’s /v1/models catalogue re-read on 2026-11-15: seven ids are gone and two joined, so every Shred tier’s stored roster drops 19 → 14 and the "+4 more" pointer is removed. Out (id absent from the catalogue): DeepSeek V4 Flash (deepseek-v4-flash-0731), Kimi K2.7 Code (kimi-k2.7-code), GLM-5.2 (glm-5.2), MiniMax M2.7 (minimax-m2.7), Qwen3.7-Max (qwen3.7-max), Qwen3.7-Plus (qwen3.7-plus) and Qwen3.7-Flash (qwen3.7-flash). In: Kimi K3.1 Pro (kimi-k3.1-pro) and Muse Spark Open (muse-spark-open), both already in the model registry and priced at Hyper’s own rows — not another provider’s (Kimi K3.1 Pro is $2.00/$10.00/$0.20 at Nous Portal and Ollama; Hyper publishes $2.20/$11.00/$0.22). Five catalog ids stay unstored: prism, prism-core and prism-flash are routing models, not selectable models — the page reads "Prism comes in three cuts, and every Hyper user has all three: pick how far up the roster each request may reach", the routing fee is "$0.10 /M Input + Cache Hit + the price of the model it picks", every plan card still lists "Prism routing model" as a feature, and the three rows reuse their default routed model’s rates, so they are not user-priced rows; gemma-4-26b-a4b-it still has no "Gemma 4 26B A4B" entry in the model registry; deepseek-v4.1-flash-1017-economy ("DeepSeek V4.1 Flash 1017 (Economy)", $0.20/$0.60/$0.01) is a distinct cheaper service tier with no registry name and no provider copy defining it, so no canonical name is invented for it. Plan prices and every quota string are unchanged.

Verified 2026-11-15 · source

Affected plans

More Charm Hyper events

Newsletter

Never miss a plan churn — get the weekly brief.

A plain weekly digest of every tracked price move, ceiling change and model-list update — pulled straight from the dated changelog ledger, with a one-line “worth switching?” note on each. Free. No spam, unsubscribe anytime.

Free · weekly · unsubscribe anytime