From 173517be27b76a38a7cc46585776e51d2175c0d4 Mon Sep 17 00:00:00 2001 From: HomeLab Agent Date: Sat, 14 Nov 2026 07:23:53 -0500 Subject: [PATCH 1/2] feat(server): offer Claude Fable 5.2 in the model picker Fable 5.2 (claude-fable-5-2, released 2026-11-09) joins the catalog above Fable 5.1: 1M context, $8/$40 per MTok (20% under 5.1), and the same 0.025x cache-read rate. Fable 5.1 and Fable 5 stay so existing pins keep resolving. The default model is unchanged (Opus 5.5). Co-Authored-By: Claude Opus 5.5 --- .changeset/offer-fable-5-2.md | 5 ++++ README.md | 2 +- packages/server/src/models.ts | 24 ++++++++++++------- .../server/test/integration/routes.test.ts | 1 + packages/server/test/unit/models.test.ts | 8 +++++++ packages/server/test/unit/usage.test.ts | 12 ++++++++++ 6 files changed, 42 insertions(+), 10 deletions(-) create mode 100644 .changeset/offer-fable-5-2.md diff --git a/.changeset/offer-fable-5-2.md b/.changeset/offer-fable-5-2.md new file mode 100644 index 00000000..fe031183 --- /dev/null +++ b/.changeset/offer-fable-5-2.md @@ -0,0 +1,5 @@ +--- +"@paddock/server": patch +--- + +Offer Claude Fable 5.2 (`claude-fable-5-2`) in the model picker. It supersedes Fable 5.1, which stays in the catalog alongside Fable 5 so existing per-project pins keep resolving. Fable 5.2 is priced at $8/$40 per million tokens (20% under Fable 5.1) and keeps Fable's 0.025Γ— cache-read rate, so a chat's estimated cost reflects the lower price. The default model is unchanged (Opus 5.5). diff --git a/README.md b/README.md index 189d451b..bb9bf117 100644 --- a/README.md +++ b/README.md @@ -77,7 +77,7 @@ you visit, more a service your other tools talk to. - πŸ“± **Works from your phone** β€” the same launchpad, fully responsive - πŸ”€ **Chat ergonomics** β€” star to pin, mark unread, fork or rewind from any message, queue-while-streaming, stop, search, archive; spawned chats nest under the chat that created them - πŸŽ›οΈ **Settings, per project and per instance** β€” model, permission mode, curation budgets and more per project; an instance-wide **Config** screen edits `paddock.config.yaml` from the UI -- 🧠 **Claude Opus 5.5 by default** β€” with Opus 5, Opus 4.8, Fable 5.1, Fable 5, Sonnet 5 and Haiku 4.5 selectable, and an allow-list if you'd rather offer fewer +- 🧠 **Claude Opus 5.5 by default** β€” with Opus 5, Opus 4.8, Fable 5.2, Fable 5.1, Fable 5, Sonnet 5 and Haiku 4.5 selectable, and an allow-list if you'd rather offer fewer - πŸ“ˆ **Token & cost tracking** β€” per-chat context meter and estimated API cost, live - πŸŽ™οΈ **Voice dictation & slash commands** β€” mic-to-text in the composer, `/`-autocomplete for skills - πŸ”Œ **Built on herdctl** β€” anything the fleet engine can do, Paddock can wire in diff --git a/packages/server/src/models.ts b/packages/server/src/models.ts index 49a5f11e..30a0c329 100644 --- a/packages/server/src/models.ts +++ b/packages/server/src/models.ts @@ -9,9 +9,9 @@ * constants β€” so changing the available models or a context limit is a one-file * edit here. * - * Context limits (verified against the Models API): Opus 5.5, Opus 5, Fable 5.1, - * Fable 5, Opus 4.8 and Sonnet 5 all have a 1,000,000-token context window; - * Haiku 4.5 is 200,000. + * Context limits (verified against the Models API): Opus 5.5, Opus 5, Fable 5.2, + * Fable 5.1, Fable 5, Opus 4.8 and Sonnet 5 all have a 1,000,000-token context + * window; Haiku 4.5 is 200,000. * On the Max plan the keeper agents run Opus at its full 1M window, so the * context meter must use 1M for it β€” otherwise a long chat reads >100%. */ @@ -32,12 +32,12 @@ export interface ModelPricing { outputPer1M: number; /** * Cache-read price as a multiple of {@link inputPer1M}, when the model departs - * from the standard 0.1Γ— ({@link CACHE_READ_MULTIPLIER}). Two models do: - * Fable 5.1 (and Mythos 5.1) read cached tokens at **0.025Γ—** ($0.25/MTok - * against a $10 base), and Opus 5.5 at **0.05Γ—** ($0.20/MTok against a $4 - * base). Every other model is the standard tenth. Cache reads dominate a long - * chat's token counts, so getting this wrong overstates a Fable 5.1 chat's - * cost ~4Γ— and an Opus 5.5 chat's ~2Γ—. + * from the standard 0.1Γ— ({@link CACHE_READ_MULTIPLIER}). Three models do: + * Fable 5.2 and Fable 5.1 (and their Mythos twins) read cached tokens at + * **0.025Γ—** ($0.20/MTok against an $8 base, $0.25 against $10), and Opus 5.5 + * at **0.05Γ—** ($0.20/MTok against a $4 base). Every other model is the + * standard tenth. Cache reads dominate a long chat's token counts, so getting + * this wrong overstates a Fable chat's cost ~4Γ— and an Opus 5.5 chat's ~2Γ—. */ cacheReadMultiplier?: number; } @@ -78,6 +78,12 @@ export const MODELS: ModelInfo[] = [ contextLimit: 1_000_000, pricing: { inputPer1M: 5, outputPer1M: 25 }, }, + { + id: "claude-fable-5-2", + label: "Fable 5.2", + contextLimit: 1_000_000, + pricing: { inputPer1M: 8, outputPer1M: 40, cacheReadMultiplier: 0.025 }, + }, { id: "claude-fable-5-1", label: "Fable 5.1", diff --git a/packages/server/test/integration/routes.test.ts b/packages/server/test/integration/routes.test.ts index 98419994..17704967 100644 --- a/packages/server/test/integration/routes.test.ts +++ b/packages/server/test/integration/routes.test.ts @@ -70,6 +70,7 @@ describe("integration: REST route coverage (real app, fake claude)", () => { expect(ids).toContain("claude-opus-5-5"); expect(ids).toContain("claude-opus-5"); expect(ids).toContain("claude-opus-4-8"); + expect(ids).toContain("claude-fable-5-2"); expect(ids).toContain("claude-fable-5-1"); expect(ids).toContain("claude-fable-5"); expect(ids).toContain("claude-sonnet-5"); diff --git a/packages/server/test/unit/models.test.ts b/packages/server/test/unit/models.test.ts index 067cedb9..345d26dd 100644 --- a/packages/server/test/unit/models.test.ts +++ b/packages/server/test/unit/models.test.ts @@ -19,6 +19,7 @@ describe("models", () => { "claude-opus-5-5", "claude-opus-5", "claude-opus-4-8", + "claude-fable-5-2", "claude-fable-5-1", "claude-fable-5", "claude-sonnet-5", @@ -61,6 +62,13 @@ describe("models", () => { contextLimit: 1_000_000, pricing: { inputPer1M: 10, outputPer1M: 50, cacheReadMultiplier: 0.025 }, }); + // Fable 5.2 is 20% cheaper than 5.1 per token but keeps the 0.025Γ— cache read. + expect(getModelInfo("claude-fable-5-2")).toEqual({ + id: "claude-fable-5-2", + label: "Fable 5.2", + contextLimit: 1_000_000, + pricing: { inputPer1M: 8, outputPer1M: 40, cacheReadMultiplier: 0.025 }, + }); // Opus 5.5 is the second model to depart from the standard 0.1Γ— cache read. expect(getModelInfo("claude-opus-5-5")).toEqual({ id: "claude-opus-5-5", diff --git a/packages/server/test/unit/usage.test.ts b/packages/server/test/unit/usage.test.ts index 7c355103..39ca9edb 100644 --- a/packages/server/test/unit/usage.test.ts +++ b/packages/server/test/unit/usage.test.ts @@ -168,6 +168,18 @@ describe("estimateCostUsd (models)", () => { ).toBeCloseTo(1.0, 6); }); + it("prices Fable 5.2 at $8/$40 with the 0.025Γ— cache read", () => { + // Fable 5.2: input $8/1M, output $40/1M β€” 20% under Fable 5.1 β€” with the + // same 0.025Γ— cache read, so $0.20/1M rather than 5.1's $0.25. + const cost = estimateCostUsd("claude-fable-5-2", { + inputTokens: 1_000_000, // $8.00 + outputTokens: 1_000_000, // $40.00 + cacheCreationTokens: 1_000_000, // $10.00 (8 Γ— 1.25) + cacheReadTokens: 1_000_000, // $0.20 (8 Γ— 0.025), NOT $0.80 + }); + expect(cost).toBeCloseTo(58.2, 6); + }); + it("honours Opus 5.5's 0.05Γ— cache-read multiplier", () => { // Opus 5.5: input $4/1M, output $20/1M; cache-read is 0.05Γ—in = $0.20/1M β€” // half the standard 0.1Γ— tenth. Second model to depart from the shared rate. From 8ece88522655d9b432e501428a3b98914414d804 Mon Sep 17 00:00:00 2001 From: HomeLab Agent Date: Sat, 14 Nov 2026 07:36:19 -0500 Subject: [PATCH 2/2] feat(server): offer Claude Sonnet 5.5 and Haiku 5.5 alongside Fable 5.2 Sonnet 5.5 (claude-sonnet-5-5): 1M context, $2/$10, 0.05x cache reads. Haiku 5.5 (claude-haiku-5-5): 1M context, $0.10/$0.50 at its base tier. Haiku 5.5 is a 1M-context model, unlike the 200K Haiku 4.5 it supersedes, so the picker hint no longer claims Haiku means 200K. Haiku 5.5 is also the one model priced by prompt length ($0.50/$2.50 over a 100k-token prompt). estimateCostUsd prices cumulative totals, which carry no per-request prompt length, so it is pinned at the base tier and a long chat is under-estimated. Documented at the catalog entry, pinned by a test, and tracked in #1003 rather than folded in here. Co-Authored-By: Claude Opus 5 --- .changeset/offer-fable-5-2.md | 4 +- README.md | 2 +- packages/server/src/models.ts | 36 +++++++++++--- .../server/test/integration/routes.test.ts | 2 + packages/server/test/unit/models.test.ts | 20 ++++++++ packages/server/test/unit/usage.test.ts | 47 +++++++++++++++++++ .../InstanceConfigStringListField.test.tsx | 2 + packages/web/src/components/SettingsPane.tsx | 2 +- 8 files changed, 105 insertions(+), 10 deletions(-) diff --git a/.changeset/offer-fable-5-2.md b/.changeset/offer-fable-5-2.md index fe031183..a266fcdc 100644 --- a/.changeset/offer-fable-5-2.md +++ b/.changeset/offer-fable-5-2.md @@ -2,4 +2,6 @@ "@paddock/server": patch --- -Offer Claude Fable 5.2 (`claude-fable-5-2`) in the model picker. It supersedes Fable 5.1, which stays in the catalog alongside Fable 5 so existing per-project pins keep resolving. Fable 5.2 is priced at $8/$40 per million tokens (20% under Fable 5.1) and keeps Fable's 0.025Γ— cache-read rate, so a chat's estimated cost reflects the lower price. The default model is unchanged (Opus 5.5). +Offer the three newest Claude models in the picker: **Fable 5.2** (`claude-fable-5-2`, $8/$40, 0.025Γ— cache reads), **Sonnet 5.5** (`claude-sonnet-5-5`, $2/$10, 0.05Γ— cache reads) and **Haiku 5.5** (`claude-haiku-5-5`, $0.10/$0.50). Each supersedes a model that stays in the catalog, so existing per-project pins keep resolving, and the default is unchanged (Opus 5.5). + +Note that Haiku 5.5 is a **1M-context** model, unlike the 200K Haiku 4.5 it supersedes β€” the model-picker hint no longer says Haiku means 200K. Its cost estimate is pinned at the base (≀100k-prompt) tier: Haiku 5.5 is the one model priced by prompt length, and a chat's cumulative token totals can't distinguish the tiers, so a long Haiku 5.5 chat is under-estimated. Tracked separately. diff --git a/README.md b/README.md index bb9bf117..98fabd20 100644 --- a/README.md +++ b/README.md @@ -77,7 +77,7 @@ you visit, more a service your other tools talk to. - πŸ“± **Works from your phone** β€” the same launchpad, fully responsive - πŸ”€ **Chat ergonomics** β€” star to pin, mark unread, fork or rewind from any message, queue-while-streaming, stop, search, archive; spawned chats nest under the chat that created them - πŸŽ›οΈ **Settings, per project and per instance** β€” model, permission mode, curation budgets and more per project; an instance-wide **Config** screen edits `paddock.config.yaml` from the UI -- 🧠 **Claude Opus 5.5 by default** β€” with Opus 5, Opus 4.8, Fable 5.2, Fable 5.1, Fable 5, Sonnet 5 and Haiku 4.5 selectable, and an allow-list if you'd rather offer fewer +- 🧠 **Claude Opus 5.5 by default** β€” with Opus 5, Opus 4.8, Fable 5.2, Fable 5.1, Fable 5, Sonnet 5.5, Sonnet 5, Haiku 5.5 and Haiku 4.5 selectable, and an allow-list if you'd rather offer fewer - πŸ“ˆ **Token & cost tracking** β€” per-chat context meter and estimated API cost, live - πŸŽ™οΈ **Voice dictation & slash commands** β€” mic-to-text in the composer, `/`-autocomplete for skills - πŸ”Œ **Built on herdctl** β€” anything the fleet engine can do, Paddock can wire in diff --git a/packages/server/src/models.ts b/packages/server/src/models.ts index 30a0c329..47dc7341 100644 --- a/packages/server/src/models.ts +++ b/packages/server/src/models.ts @@ -9,9 +9,9 @@ * constants β€” so changing the available models or a context limit is a one-file * edit here. * - * Context limits (verified against the Models API): Opus 5.5, Opus 5, Fable 5.2, - * Fable 5.1, Fable 5, Opus 4.8 and Sonnet 5 all have a 1,000,000-token context - * window; Haiku 4.5 is 200,000. + * Context limits (verified against the Models API): every catalog model has a + * 1,000,000-token context window EXCEPT Haiku 4.5, which is 200,000. Note that + * Haiku 5.5 β€” unlike the 4.5 it supersedes β€” is a 1M model. * On the Max plan the keeper agents run Opus at its full 1M window, so the * context meter must use 1M for it β€” otherwise a long chat reads >100%. */ @@ -32,12 +32,13 @@ export interface ModelPricing { outputPer1M: number; /** * Cache-read price as a multiple of {@link inputPer1M}, when the model departs - * from the standard 0.1Γ— ({@link CACHE_READ_MULTIPLIER}). Three models do: + * from the standard 0.1Γ— ({@link CACHE_READ_MULTIPLIER}). Four models do: * Fable 5.2 and Fable 5.1 (and their Mythos twins) read cached tokens at * **0.025Γ—** ($0.20/MTok against an $8 base, $0.25 against $10), and Opus 5.5 - * at **0.05Γ—** ($0.20/MTok against a $4 base). Every other model is the - * standard tenth. Cache reads dominate a long chat's token counts, so getting - * this wrong overstates a Fable chat's cost ~4Γ— and an Opus 5.5 chat's ~2Γ—. + * and Sonnet 5.5 at **0.05Γ—** ($0.20/MTok against a $4 base, $0.10 against + * $2). Every other model is the standard tenth. Cache reads dominate a long + * chat's token counts, so getting this wrong overstates a Fable chat's cost + * ~4Γ— and an Opus 5.5 / Sonnet 5.5 chat's ~2Γ—. */ cacheReadMultiplier?: number; } @@ -96,6 +97,12 @@ export const MODELS: ModelInfo[] = [ contextLimit: 1_000_000, pricing: { inputPer1M: 10, outputPer1M: 50 }, }, + { + id: "claude-sonnet-5-5", + label: "Sonnet 5.5", + contextLimit: 1_000_000, + pricing: { inputPer1M: 2, outputPer1M: 10, cacheReadMultiplier: 0.05 }, + }, { id: "claude-sonnet-5", label: "Sonnet 5", @@ -106,6 +113,21 @@ export const MODELS: ModelInfo[] = [ // happened, overstating every Sonnet 5 cost estimate by 50%. pricing: { inputPer1M: 2, outputPer1M: 10 }, }, + { + id: "claude-haiku-5-5", + label: "Haiku 5.5", + contextLimit: 1_000_000, + // ⚠️ Haiku 5.5 is the one model priced by PROMPT LENGTH, not by a flat rate: + // a request whose prompt (input + cache reads + cache writes) exceeds + // 100,000 tokens is billed at $0.50/$2.50 rather than the $0.10/$0.50 below, + // each request judged on its own. {@link estimateCostUsd} prices a chat's + // CUMULATIVE totals, which no longer carry per-request prompt lengths, so it + // cannot tell the tiers apart and always applies the cheaper one. A chat that + // runs long on Haiku 5.5 is therefore UNDER-estimated, by up to 5Γ—. Priced at + // the base tier deliberately: it is exact for short chats (the common case for + // the cheap model) rather than wrong in both directions. Tracked in #1003. + pricing: { inputPer1M: 0.1, outputPer1M: 0.5 }, + }, { id: "claude-haiku-4-5-20251001", label: "Haiku 4.5", diff --git a/packages/server/test/integration/routes.test.ts b/packages/server/test/integration/routes.test.ts index 17704967..fb704e49 100644 --- a/packages/server/test/integration/routes.test.ts +++ b/packages/server/test/integration/routes.test.ts @@ -73,7 +73,9 @@ describe("integration: REST route coverage (real app, fake claude)", () => { expect(ids).toContain("claude-fable-5-2"); expect(ids).toContain("claude-fable-5-1"); expect(ids).toContain("claude-fable-5"); + expect(ids).toContain("claude-sonnet-5-5"); expect(ids).toContain("claude-sonnet-5"); + expect(ids).toContain("claude-haiku-5-5"); // Every model carries a positive context limit (drives the meter). for (const m of body.models) expect(m.contextLimit).toBeGreaterThan(0); // The box-wide keeper drive-mode default (issue #122): the Settings tab reads diff --git a/packages/server/test/unit/models.test.ts b/packages/server/test/unit/models.test.ts index 345d26dd..008b5ed1 100644 --- a/packages/server/test/unit/models.test.ts +++ b/packages/server/test/unit/models.test.ts @@ -22,7 +22,9 @@ describe("models", () => { "claude-fable-5-2", "claude-fable-5-1", "claude-fable-5", + "claude-sonnet-5-5", "claude-sonnet-5", + "claude-haiku-5-5", "claude-haiku-4-5-20251001", ]); expect(MODELS[0].id).toBe(DEFAULT_MODEL); @@ -46,6 +48,9 @@ describe("models", () => { expect(getContextLimit("claude-opus-5")).toBe(1_000_000); expect(getContextLimit("claude-opus-4-8")).toBe(1_000_000); expect(getContextLimit("claude-haiku-4-5-20251001")).toBe(200_000); + // Haiku 5.5 breaks the "Haiku means 200k" habit β€” it is a 1M model, and the + // picker hint in SettingsPane.tsx is worded to match. + expect(getContextLimit("claude-haiku-5-5")).toBe(1_000_000); expect(getContextLimit("nope")).toBe(200_000); }); @@ -76,6 +81,21 @@ describe("models", () => { contextLimit: 1_000_000, pricing: { inputPer1M: 4, outputPer1M: 20, cacheReadMultiplier: 0.05 }, }); + expect(getModelInfo("claude-sonnet-5-5")).toEqual({ + id: "claude-sonnet-5-5", + label: "Sonnet 5.5", + contextLimit: 1_000_000, + pricing: { inputPer1M: 2, outputPer1M: 10, cacheReadMultiplier: 0.05 }, + }); + // Haiku 5.5 is pinned at its BASE (≀100k prompt) tier. The over-100k tier + // ($0.50/$2.50) is deliberately not modelled β€” see the note in models.ts and + // issue #1003 β€” so this is the rate a long chat is under-estimated at. + expect(getModelInfo("claude-haiku-5-5")).toEqual({ + id: "claude-haiku-5-5", + label: "Haiku 5.5", + contextLimit: 1_000_000, + pricing: { inputPer1M: 0.1, outputPer1M: 0.5 }, + }); // Pinned because we shipped $3/$15 for a price rise that was cancelled; the // standard price is $2/$10. Guards against re-importing the stale schedule. expect(getModelInfo("claude-sonnet-5")).toEqual({ diff --git a/packages/server/test/unit/usage.test.ts b/packages/server/test/unit/usage.test.ts index 39ca9edb..537d7ba2 100644 --- a/packages/server/test/unit/usage.test.ts +++ b/packages/server/test/unit/usage.test.ts @@ -180,6 +180,53 @@ describe("estimateCostUsd (models)", () => { expect(cost).toBeCloseTo(58.2, 6); }); + it("prices Sonnet 5.5 at $2/$10 with the 0.05Γ— cache read", () => { + // Sonnet 5.5 keeps Sonnet 5's $2/$10 but halves the cache read: 0.05Γ—in = + // $0.10/1M, where Sonnet 5 pays the standard $0.20. + const cost = estimateCostUsd("claude-sonnet-5-5", { + inputTokens: 0, + outputTokens: 0, + cacheCreationTokens: 0, + cacheReadTokens: 1_000_000, // $0.10 (2 Γ— 0.05), NOT $0.20 + }); + expect(cost).toBeCloseTo(0.1, 6); + + // Control: Sonnet 5 has no override, so it stays on the standard tenth. + expect( + estimateCostUsd("claude-sonnet-5", { + inputTokens: 0, + outputTokens: 0, + cacheCreationTokens: 0, + cacheReadTokens: 1_000_000, + }), + ).toBeCloseTo(0.2, 6); + }); + + it("prices Haiku 5.5 at its base tier, flat β€” the documented long-prompt gap", () => { + // Base (≀100k prompt) tier: input $0.10/1M, output $0.50/1M, cache write + // 1.25Γ—in = $0.125, cache read the standard 0.1Γ—in = $0.01. + const cost = estimateCostUsd("claude-haiku-5-5", { + inputTokens: 1_000_000, // $0.10 + outputTokens: 1_000_000, // $0.50 + cacheCreationTokens: 1_000_000, // $0.125 + cacheReadTokens: 1_000_000, // $0.01 + }); + expect(cost).toBeCloseTo(0.735, 6); + + // KNOWN LIMITATION (#1003), pinned so a change is deliberate: Haiku 5.5 bills + // a request whose prompt exceeds 100k tokens at $0.50/$2.50 instead, but + // cumulative totals carry no per-request prompt length, so the estimate stays + // on the cheap tier however large the chat gets β€” scaling here is linear, not + // tiered. A long Haiku 5.5 chat is under-estimated by up to 5Γ—. + const big = estimateCostUsd("claude-haiku-5-5", { + inputTokens: 10_000_000, + outputTokens: 10_000_000, + cacheCreationTokens: 10_000_000, + cacheReadTokens: 10_000_000, + }); + expect(big).toBeCloseTo(7.35, 6); + }); + it("honours Opus 5.5's 0.05Γ— cache-read multiplier", () => { // Opus 5.5: input $4/1M, output $20/1M; cache-read is 0.05Γ—in = $0.20/1M β€” // half the standard 0.1Γ— tenth. Second model to depart from the shared rate. diff --git a/packages/web/src/components/InstanceConfigStringListField.test.tsx b/packages/web/src/components/InstanceConfigStringListField.test.tsx index f40f036a..19a37641 100644 --- a/packages/web/src/components/InstanceConfigStringListField.test.tsx +++ b/packages/web/src/components/InstanceConfigStringListField.test.tsx @@ -40,7 +40,9 @@ const CATALOG = [ "claude-opus-4-8", "claude-fable-5-1", "claude-fable-5", + "claude-sonnet-5-5", "claude-sonnet-5", + "claude-haiku-5-5", "claude-haiku-4-5-20251001", ]; const JOINED = CATALOG.join(", "); diff --git a/packages/web/src/components/SettingsPane.tsx b/packages/web/src/components/SettingsPane.tsx index 96d32875..4395a63f 100644 --- a/packages/web/src/components/SettingsPane.tsx +++ b/packages/web/src/components/SettingsPane.tsx @@ -539,7 +539,7 @@ export function SettingsPane({ ))} - Larger context windows (Opus/Fable/Sonnet: 1M; Haiku: 200K) fit longer chats. + Larger context windows fit longer chats β€” every model here is 1M except Haiku 4.5 (200K).