From e818dc80e5ab86b4704481727f3b71b11380b73e Mon Sep 17 00:00:00 2001
From: Waleed Latif <walif6@gmail.com>
Date: Mon, 28 Sep 2026 12:39:31 -0700
Subject: [PATCH 1/3] feat(providers): add Claude Sonnet 5.5 and make it the
 default model

---
 apps/docs/content/docs/agents/index.mdx       |  2 +-
 .../content/docs/getting-started/index.mdx    |  2 +-
 .../content/docs/workflows/blocks/agent.mdx   |  4 +-
 .../docs/workflows/blocks/evaluator.mdx       |  2 +-
 .../content/docs/workflows/blocks/router.mdx  |  2 +-
 .../components/hero-visual/workflow-data.ts   |  2 +-
 .../composer-loop/demo-workflow-data.ts       |  4 +-
 .../components/combobox/combobox.tsx          |  5 +-
 apps/sim/blocks/blocks/agent.ts               |  4 +-
 apps/sim/blocks/blocks/evaluator.ts           |  2 +-
 apps/sim/blocks/blocks/router.ts              |  4 +-
 apps/sim/executor/constants.ts                |  6 +--
 apps/sim/providers/anthropic/core.test.ts     | 54 ++++++++++++++++++-
 apps/sim/providers/anthropic/core.ts          | 41 ++++++++++----
 apps/sim/providers/models.ts                  | 34 +++++++++++-
 15 files changed, 135 insertions(+), 33 deletions(-)

diff --git a/apps/docs/content/docs/agents/index.mdx b/apps/docs/content/docs/agents/index.mdx
index 7ab7f7afc70..2fe27a3073c 100644
--- a/apps/docs/content/docs/agents/index.mdx
+++ b/apps/docs/content/docs/agents/index.mdx
@@ -17,7 +17,7 @@ The example throughout is an agent that scores inbound sales leads.
 
 ## The Agent block
 
-You set up the reasoning step by giving the Agent block a **model** and a **prompt**. The model is the LLM that powers it; you pick one from the available providers, and the default is `claude-sonnet-5`. The prompt is a system message that defines who the agent is and how it should behave, plus a user message that carries the input, usually a reference like `<start.input>`.
+You set up the reasoning step by giving the Agent block a **model** and a **prompt**. The model is the LLM that powers it; you pick one from the available providers, and the default is `claude-sonnet-5-5`. The prompt is a system message that defines who the agent is and how it should behave, plus a user message that carries the input, usually a reference like `<start.input>`.
 
 When it runs, the Agent block reasons, calls any tools it needs, and stores its result under its own name. By default that result is free text in `content`, read by a later block as `<agent.content>`, alongside run details like the model used, token counts, tool calls, and cost. Every setting and output field is in the [Agent block reference](/workflows/blocks/agent).
 
diff --git a/apps/docs/content/docs/getting-started/index.mdx b/apps/docs/content/docs/getting-started/index.mdx
index 99737fa5f73..4156a12b51b 100644
--- a/apps/docs/content/docs/getting-started/index.mdx
+++ b/apps/docs/content/docs/getting-started/index.mdx
@@ -28,7 +28,7 @@ Build a people research agent in 10 minutes. It takes a name through a chat inte
     - **System**: "You are a people research agent. When given a person's name, use your search tools to find their location, profession, educational background, and other relevant details."
     - **User**: insert `<start.input>` so the agent reads whatever the chat receives.
 
-    Leave the **Model** on the default (`claude-sonnet-5`), or pick any other.
+    Leave the **Model** on the default (`claude-sonnet-5-5`), or pick any other.
 
     <div className="mx-auto w-full overflow-hidden rounded-lg">
       <Video src="getting-started/started-2.mp4" width={700} height={450} />
diff --git a/apps/docs/content/docs/workflows/blocks/agent.mdx b/apps/docs/content/docs/workflows/blocks/agent.mdx
index f754555a240..a07a9aa4658 100644
--- a/apps/docs/content/docs/workflows/blocks/agent.mdx
+++ b/apps/docs/content/docs/workflows/blocks/agent.mdx
@@ -25,7 +25,7 @@ Answer in two sentences, cite the doc you used, and never guess a price.
 
 ### Model
 
-The model that runs the step. Defaults to `claude-sonnet-5`. Type or pick any model from OpenAI, Anthropic, Google, xAI, Groq, Cerebras, DeepSeek, Azure, AWS Bedrock, Google Vertex, or OpenRouter, or a local model through Ollama or VLLM.
+The model that runs the step. Defaults to `claude-sonnet-5-5`. Type or pick any model from OpenAI, Anthropic, Google, xAI, Groq, Cerebras, DeepSeek, Azure, AWS Bedrock, Google Vertex, or OpenRouter, or a local model through Ollama or VLLM.
 
 For a custom cloud deployment, enter its provider prefix and model ID: `azure/my-deployment`, `azure-anthropic/my-deployment`, `bedrock/my-inference-profile`, or `vertex/my-gemini-model`. The prefix selects the provider and shows its credential fields even when the ID is absent from the catalog. Bedrock accepts full inference profile ARNs after `bedrock/`; Vertex uses the Gemini API and accepts Google model resource names. The deployment must support the selected provider's API. Custom IDs have no catalog pricing or token limits.
 
@@ -159,7 +159,7 @@ Live tool-call chips stream for **OpenAI, Anthropic, Azure Anthropic, Google, Ve
 | Provider | Streamed thinking | Models |
 |----------|-------------------|--------|
 | OpenAI | Summaries only — Requires OpenAI organization verification; falls back to no summaries. | `gpt-6-astra`, `gpt-6-sol`, `gpt-6-luna`, `gpt-5.6-sol`, `gpt-5.6-terra`, `gpt-5.6-luna`, `gpt-5.5-pro`, `gpt-5.5`, `gpt-5.4-pro`, `gpt-5.4`, `gpt-5.4-mini`, `gpt-5.4-nano`, `gpt-5.3-codex`, `gpt-5.2-pro`, `gpt-5.2`, `gpt-5.1`, `gpt-5-pro`, `gpt-5`, `gpt-5-mini`, `gpt-5-nano`, `o4-mini`, `o3`, `o3-mini`, `o1` |
-| Anthropic | Summaries only — These generations omit full thinking; Sim requests summarized thinking on streaming runs. | `claude-fable-5-1`, `claude-fable-5`, `claude-sonnet-5`, `claude-opus-5-5`, `claude-opus-5`, `claude-opus-4-8`, `claude-opus-4-7`, `claude-opus-4-6`, `claude-sonnet-4-6`, `claude-opus-4-5`, `claude-sonnet-4-5`, `claude-haiku-4-5` |
+| Anthropic | Summaries only — These generations omit full thinking; Sim requests summarized thinking on streaming runs. | `claude-fable-5-1`, `claude-fable-5`, `claude-sonnet-5-5`, `claude-sonnet-5`, `claude-opus-5-5`, `claude-opus-5`, `claude-opus-4-8`, `claude-opus-4-7`, `claude-opus-4-6`, `claude-sonnet-4-6`, `claude-opus-4-5`, `claude-sonnet-4-5`, `claude-haiku-4-5` |
 | Azure OpenAI | Summaries only — Requires OpenAI organization verification; falls back to no summaries. | `azure/gpt-6-astra`, `azure/gpt-5.6-sol`, `azure/gpt-5.6-terra`, `azure/gpt-5.6-luna`, `azure/gpt-5.5`, `azure/gpt-5.4-pro`, `azure/gpt-5.4`, `azure/gpt-5.4-mini`, `azure/gpt-5.4-nano`, `azure/gpt-5.2`, `azure/gpt-5.1`, `azure/gpt-5.1-codex`, `azure/gpt-5`, `azure/gpt-5-mini`, `azure/gpt-5-nano`, `azure/o3`, `azure/o4-mini` |
 | Azure Anthropic | Summaries only — These generations omit full thinking; Sim requests summarized thinking on streaming runs. | `azure-anthropic/claude-fable-5-1`, `azure-anthropic/claude-opus-5`, `azure-anthropic/claude-opus-4-8`, `azure-anthropic/claude-opus-4-7`, `azure-anthropic/claude-opus-4-6`, `azure-anthropic/claude-opus-4-5`, `azure-anthropic/claude-sonnet-5`, `azure-anthropic/claude-sonnet-4-6`, `azure-anthropic/claude-sonnet-4-5`, `azure-anthropic/claude-opus-4-1`, `azure-anthropic/claude-haiku-4-5` |
 | Google | Summaries only | `gemini-3.8-flash`, `gemini-3.7-flash`, `gemini-3.6-flash`, `gemini-3.5-flash-lite`, `gemini-3.5-flash`, `gemini-3.1-pro-preview`, `gemini-3.1-flash-lite`, `gemini-3-flash-preview`, `gemini-2.5-pro`, `gemini-2.5-flash`, `gemini-2.5-flash-lite` |
diff --git a/apps/docs/content/docs/workflows/blocks/evaluator.mdx b/apps/docs/content/docs/workflows/blocks/evaluator.mdx
index eb21db96c74..f99593a0800 100644
--- a/apps/docs/content/docs/workflows/blocks/evaluator.mdx
+++ b/apps/docs/content/docs/workflows/blocks/evaluator.mdx
@@ -30,7 +30,7 @@ The content to score. Usually an earlier output like `<agent.content>`. Structur
 
 ### Model
 
-The model that does the scoring, defaulting to `claude-sonnet-5`. Stronger reasoning models give more consistent scores. Type or pick any supported model. **Temperature** and a **System Prompt** are available under advanced, and on hosted Sim the API key is supplied for you.
+The model that does the scoring, defaulting to `claude-sonnet-5-5`. Stronger reasoning models give more consistent scores. Type or pick any supported model. **Temperature** and a **System Prompt** are available under advanced, and on hosted Sim the API key is supplied for you.
 
 ### Fallback models
 
diff --git a/apps/docs/content/docs/workflows/blocks/router.mdx b/apps/docs/content/docs/workflows/blocks/router.mdx
index 0cc2b8fa2bb..926a9e62799 100644
--- a/apps/docs/content/docs/workflows/blocks/router.mdx
+++ b/apps/docs/content/docs/workflows/blocks/router.mdx
@@ -33,7 +33,7 @@ Each route is a **title** and a **description** of when to choose it ("Route her
 
 ### Model
 
-The model that makes the decision, defaulting to `claude-sonnet-5`. Stronger reasoning models route more accurately; a faster, cheaper model is fine when the routes are clearly distinct. Type or pick any supported model, or a local one through Ollama or VLLM. On hosted Sim the API key is supplied for you.
+The model that makes the decision, defaulting to `claude-sonnet-5-5`. Stronger reasoning models route more accurately; a faster, cheaper model is fine when the routes are clearly distinct. Type or pick any supported model, or a local one through Ollama or VLLM. On hosted Sim the API key is supplied for you.
 
 ### Fallback models
 
diff --git a/apps/sim/app/(landing)/components/hero/components/hero-visual/workflow-data.ts b/apps/sim/app/(landing)/components/hero/components/hero-visual/workflow-data.ts
index 29bc4102838..d8a29d1dafb 100644
--- a/apps/sim/app/(landing)/components/hero/components/hero-visual/workflow-data.ts
+++ b/apps/sim/app/(landing)/components/hero/components/hero-visual/workflow-data.ts
@@ -162,7 +162,7 @@ export const BLOCKS: BlockDef[] = [
     bgColor: 'var(--text-primary)',
     sentence: {
       segments: ['Prompt', { subBlockId: 'model', noun: 'a model' }],
-      values: { model: 'claude-sonnet-5' },
+      values: { model: 'claude-sonnet-5-5' },
     },
     rows: [],
     x: 300,
diff --git a/apps/sim/app/(landing)/components/product-demo/components/composer-loop/demo-workflow-data.ts b/apps/sim/app/(landing)/components/product-demo/components/composer-loop/demo-workflow-data.ts
index 9feed6b82e3..1cda204fc52 100644
--- a/apps/sim/app/(landing)/components/product-demo/components/composer-loop/demo-workflow-data.ts
+++ b/apps/sim/app/(landing)/components/product-demo/components/composer-loop/demo-workflow-data.ts
@@ -98,7 +98,7 @@ export const DEMO_BLOCKS: BlockDef[] = [
     bgColor: 'var(--text-primary)',
     sentence: {
       segments: ['Prompt', { subBlockId: 'model', noun: 'a model' }],
-      values: { model: 'claude-sonnet-5' },
+      values: { model: 'claude-sonnet-5-5' },
     },
     rows: [],
     x: col(2),
@@ -127,7 +127,7 @@ export const DEMO_BLOCKS: BlockDef[] = [
     bgColor: 'var(--text-primary)',
     sentence: {
       segments: ['Prompt', { subBlockId: 'model', noun: 'a model' }],
-      values: { model: 'claude-sonnet-5' },
+      values: { model: 'claude-sonnet-5-5' },
     },
     rows: [],
     x: col(4),
diff --git a/apps/sim/app/workspace/[workspaceId]/w/[workflowId]/components/panel/components/editor/components/sub-block/components/combobox/combobox.tsx b/apps/sim/app/workspace/[workspaceId]/w/[workflowId]/components/panel/components/editor/components/sub-block/components/combobox/combobox.tsx
index 7b40f491164..6872c9f9893 100644
--- a/apps/sim/app/workspace/[workspaceId]/w/[workflowId]/components/panel/components/editor/components/sub-block/components/combobox/combobox.tsx
+++ b/apps/sim/app/workspace/[workspaceId]/w/[workflowId]/components/panel/components/editor/components/sub-block/components/combobox/combobox.tsx
@@ -25,7 +25,7 @@ import { useSubBlockStore } from '@/stores/workflows/subblock/store'
 /**
  * Constants for ComboBox component behavior
  */
-const DEFAULT_MODEL = 'claude-sonnet-5'
+const DEFAULT_MODEL = 'claude-sonnet-5-5'
 const ZOOM_FACTOR_BASE = 0.96
 const MIN_ZOOM = 0.1
 const MAX_ZOOM = 1
@@ -296,7 +296,7 @@ export const ComboBox = memo(function ComboBox({
 
   /**
    * Determines the default option value to use.
-   * Priority: explicit defaultValue > claude-sonnet-5 for model field > first option
+   * Priority: explicit defaultValue > DEFAULT_MODEL for model field > first option
    */
   const defaultOptionValue = useMemo(() => {
     if (defaultValue !== undefined) {
@@ -308,7 +308,6 @@ export const ComboBox = memo(function ComboBox({
       // Default not available (e.g. provider disabled) — fall through to other fallbacks
     }
 
-    // For model field, default to claude-sonnet-5 if available
     if (subBlockId === 'model') {
       const defaultModelOption = evaluatedOptions.find(
         (opt) => getOptionValue(opt) === DEFAULT_MODEL
diff --git a/apps/sim/blocks/blocks/agent.ts b/apps/sim/blocks/blocks/agent.ts
index 6cc01283268..9eb4e2732af 100644
--- a/apps/sim/blocks/blocks/agent.ts
+++ b/apps/sim/blocks/blocks/agent.ts
@@ -34,7 +34,7 @@ import type { ToolResponse } from '@/tools/types'
 const logger = createLogger('AgentBlock')
 
 /** Model the agent block falls back to when `model` is unset or the auto pseudo-model. */
-const AGENT_FALLBACK_MODEL = 'claude-sonnet-5'
+const AGENT_FALLBACK_MODEL = 'claude-sonnet-5-5'
 
 const MODELS_WITH_REASONING_EFFORT = getModelsWithReasoningEffort()
 const MODELS_WITH_VERBOSITY = getModelsWithVerbosity()
@@ -154,7 +154,7 @@ Return ONLY the JSON array.`,
       type: 'combobox',
       placeholder: 'Type or select a model...',
       required: true,
-      defaultValue: 'claude-sonnet-5',
+      defaultValue: 'claude-sonnet-5-5',
       options: getAgentModelOptions,
       commandSearchable: true,
     },
diff --git a/apps/sim/blocks/blocks/evaluator.ts b/apps/sim/blocks/blocks/evaluator.ts
index 75fed5e1c06..0d04defb16b 100644
--- a/apps/sim/blocks/blocks/evaluator.ts
+++ b/apps/sim/blocks/blocks/evaluator.ts
@@ -186,7 +186,7 @@ export const EvaluatorBlock: BlockConfig<EvaluatorResponse> = {
       type: 'combobox',
       placeholder: 'Type or select a model...',
       required: true,
-      defaultValue: 'claude-sonnet-5',
+      defaultValue: 'claude-sonnet-5-5',
       options: getModelOptions,
     },
     ...getProviderCredentialSubBlocks(),
diff --git a/apps/sim/blocks/blocks/router.ts b/apps/sim/blocks/blocks/router.ts
index 8d9bccb9f9c..63e988e70e3 100644
--- a/apps/sim/blocks/blocks/router.ts
+++ b/apps/sim/blocks/blocks/router.ts
@@ -183,7 +183,7 @@ export const RouterBlock: BlockConfig<RouterResponse> = {
       type: 'combobox',
       placeholder: 'Type or select a model...',
       required: true,
-      defaultValue: 'claude-sonnet-5',
+      defaultValue: 'claude-sonnet-5-5',
       options: getModelOptions,
     },
     ...getProviderCredentialSubBlocks(),
@@ -302,7 +302,7 @@ export const RouterV2Block: BlockConfig<RouterV2Response> = {
       type: 'combobox',
       placeholder: 'Type or select a model...',
       required: true,
-      defaultValue: 'claude-sonnet-5',
+      defaultValue: 'claude-sonnet-5-5',
       options: getModelOptions,
     },
     ...getProviderCredentialSubBlocks(),
diff --git a/apps/sim/executor/constants.ts b/apps/sim/executor/constants.ts
index e0b4b08c283..00d7c9c20db 100644
--- a/apps/sim/executor/constants.ts
+++ b/apps/sim/executor/constants.ts
@@ -228,7 +228,7 @@ export const HTTP = {
 } as const
 
 export const AGENT = {
-  DEFAULT_MODEL: 'claude-sonnet-5',
+  DEFAULT_MODEL: 'claude-sonnet-5-5',
   get DEFAULT_FUNCTION_TIMEOUT() {
     return getMaxExecutionTimeout()
   },
@@ -243,13 +243,13 @@ export const MCP = {
 } as const
 
 export const ROUTER = {
-  DEFAULT_MODEL: 'claude-sonnet-5',
+  DEFAULT_MODEL: 'claude-sonnet-5-5',
   DEFAULT_TEMPERATURE: 0,
   INFERENCE_TEMPERATURE: 0.1,
 } as const
 
 export const EVALUATOR = {
-  DEFAULT_MODEL: 'claude-sonnet-5',
+  DEFAULT_MODEL: 'claude-sonnet-5-5',
   DEFAULT_TEMPERATURE: 0.1,
   RESPONSE_SCHEMA_NAME: 'evaluation_response',
   JSON_INDENT: 2,
diff --git a/apps/sim/providers/anthropic/core.test.ts b/apps/sim/providers/anthropic/core.test.ts
index 62f361f4c9b..c8f9da76f21 100644
--- a/apps/sim/providers/anthropic/core.test.ts
+++ b/apps/sim/providers/anthropic/core.test.ts
@@ -293,7 +293,7 @@ describe('executeAnthropicProviderRequest forced tool use', () => {
     expect(payload.tool_choice).toEqual({ type: 'tool', name: 'publish' })
   })
 
-  it.each(['claude-fable-5-1', 'claude-opus-5-5'])(
+  it.each(['claude-fable-5-1', 'claude-opus-5-5', 'claude-sonnet-5-5'])(
     'drops forced tool_choice when the catalog model disables Force (%s)',
     async (model) => {
       const { payload, warn } = await runWithForcedTool(model)
@@ -304,6 +304,57 @@ describe('executeAnthropicProviderRequest forced tool use', () => {
   )
 })
 
+/**
+ * Claude Sonnet 5.5 rejects `thinking.type: "disabled"` with a 400 and names `between_tools`
+ * (which takes no effort or display field) as its lowest setting. Every other model keeps
+ * `none` as "send no thinking config".
+ */
+describe('executeAnthropicProviderRequest none thinking level', () => {
+  async function sendWithThinkingNone(model: string) {
+    const create = vi.fn().mockResolvedValue({
+      id: 'msg-none',
+      type: 'message',
+      role: 'assistant',
+      model,
+      content: [{ type: 'text', text: 'Done' }],
+      stop_reason: 'end_turn',
+      stop_sequence: null,
+      usage: { input_tokens: 2, output_tokens: 2 },
+    })
+    await executeAnthropicProviderRequest(
+      {
+        model,
+        apiKey: 'test-key',
+        maxTokens: 1024,
+        thinkingLevel: 'none',
+        agentEvents: true,
+        messages: [{ role: 'user', content: 'Hello' }],
+      },
+      {
+        providerId: 'anthropic',
+        providerLabel: 'Anthropic',
+        createClient: () => ({ messages: { create } }) as never,
+        logger: { info: vi.fn(), warn: vi.fn(), error: vi.fn(), debug: vi.fn() },
+      }
+    )
+    return create.mock.calls[0][0] as Anthropic.Messages.MessageCreateParams
+  }
+
+  it('sends bare between_tools thinking on claude-sonnet-5-5', async () => {
+    const payload = await sendWithThinkingNone('claude-sonnet-5-5')
+    expect(payload.thinking).toEqual({ type: 'between_tools' })
+    expect(payload.output_config).toBeUndefined()
+  })
+
+  it.each(['claude-sonnet-5', 'claude-opus-5-5'])(
+    'sends no thinking config on %s',
+    async (model) => {
+      const payload = await sendWithThinkingNone(model)
+      expect(payload.thinking).toBeUndefined()
+    }
+  )
+})
+
 describe('executeAnthropicProviderRequest native structured outputs', () => {
   /** The subset of the wire schema these assertions read back. */
   interface WireSchemaNode {
@@ -758,6 +809,7 @@ describe('buildThinkingConfig', () => {
     for (const model of [
       'claude-fable-5-1',
       'claude-fable-5',
+      'claude-sonnet-5-5',
       'claude-sonnet-5',
       'claude-opus-5-5',
       'claude-opus-5',
diff --git a/apps/sim/providers/anthropic/core.ts b/apps/sim/providers/anthropic/core.ts
index 0f84ef8c615..3e142c6ad73 100644
--- a/apps/sim/providers/anthropic/core.ts
+++ b/apps/sim/providers/anthropic/core.ts
@@ -144,6 +144,7 @@ const ANTHROPIC_THINKING_OUTPUT_HEADROOM = 4096
 /**
  * Checks if a model supports adaptive thinking (thinking.type: "adaptive").
  * Fable 5, Fable 5.1, and Opus 5.5 support ONLY adaptive thinking (always on; type: "disabled" is rejected).
+ * Sonnet 5.5 is adaptive by default and rejects type: "disabled"; its lowest setting is "between_tools".
  * Sonnet 5 supports ONLY adaptive thinking (manual budget_tokens returns a 400 error).
  * Opus 5, Opus 4.8, and Opus 4.7 support ONLY adaptive thinking (no extended thinking / budget_tokens).
  * Opus 4.6 and Sonnet 4.6 support both extended and adaptive thinking — use adaptive.
@@ -169,10 +170,15 @@ function supportsAdaptiveThinking(modelId: string): boolean {
 /**
  * Builds the thinking configuration for the Anthropic API based on model capabilities and level.
  *
- * - Fable 5.1, Fable 5, Sonnet 5, Opus 5.5, Opus 5, Opus 4.8, Opus 4.7: Uses adaptive thinking only (no extended thinking support)
+ * - Fable 5.1, Fable 5, Sonnet 5.5, Sonnet 5, Opus 5.5, Opus 5, Opus 4.8, Opus 4.7: Uses adaptive thinking only (no extended thinking support)
  * - Opus 4.6, Sonnet 4.6: Uses adaptive thinking with effort parameter
  * - Other models: Uses budget_tokens-based extended thinking
  *
+ * The `none` level returns null (send no thinking config) unless the model
+ * declares `capabilities.thinking.noneMode`: Sonnet 5.5 rejects
+ * `type: "disabled"`, so `none` becomes `type: "between_tools"`, which turns
+ * off up-front thinking and takes no effort or display field.
+ *
  * The newest Claude generations default `thinking.display` to `omitted`
  * (empty thinking blocks, no thinking deltas). Their registry entries mark
  * `capabilities.thinking.streamed: 'summary'`, and for those models Sim opts
@@ -190,7 +196,19 @@ export function buildThinkingConfig(
   outputConfig?: Anthropic.Messages.OutputConfig
 } | null {
   const capability = getThinkingCapability(modelId)
-  if (!capability || !capability.levels.includes(thinkingLevel)) {
+  if (!capability) {
+    return null
+  }
+
+  if (thinkingLevel === 'none') {
+    if (capability.noneMode !== 'between_tools') return null
+    return {
+      // double-cast-allowed: @anthropic-ai/sdk 0.115 predates the between_tools thinking type (typed from 0.129)
+      thinking: { type: 'between_tools' } as unknown as Anthropic.Messages.ThinkingConfigParam,
+    }
+  }
+
+  if (!capability.levels.includes(thinkingLevel)) {
     return null
   }
 
@@ -358,9 +376,8 @@ export async function executeAnthropicProviderRequest(
     }
   }
 
-  // Add extended thinking configuration if supported and requested
-  // The 'none' sentinel means "disable thinking" — skip configuration entirely.
-  if (request.thinkingLevel && request.thinkingLevel !== 'none') {
+  // The 'none' sentinel means "disable thinking": no config, unless the model declares a noneMode.
+  if (request.thinkingLevel) {
     const thinkingConfig = buildThinkingConfig(
       request.model,
       request.thinkingLevel,
@@ -405,11 +422,15 @@ export async function executeAnthropicProviderRequest(
       // Per Anthropic docs: thinking is not compatible with temperature or top_k modifications.
       payload.temperature = undefined
 
-      const isAdaptive = thinkingConfig.thinking.type === 'adaptive'
-      logger.info(
-        `Using ${isAdaptive ? 'adaptive' : 'extended'} thinking for model: ${modelId} with ${isAdaptive ? `effort: ${request.thinkingLevel}` : `budget: ${(thinkingConfig.thinking as { budget_tokens: number }).budget_tokens}`}`
-      )
-    } else {
+      if (request.thinkingLevel === 'none') {
+        logger.info(`Using between_tools thinking for model: ${modelId}`)
+      } else {
+        const isAdaptive = thinkingConfig.thinking.type === 'adaptive'
+        logger.info(
+          `Using ${isAdaptive ? 'adaptive' : 'extended'} thinking for model: ${modelId} with ${isAdaptive ? `effort: ${request.thinkingLevel}` : `budget: ${(thinkingConfig.thinking as { budget_tokens: number }).budget_tokens}`}`
+        )
+      }
+    } else if (request.thinkingLevel !== 'none') {
       logger.warn(
         `Thinking level "${describeModelLevel(request.thinkingLevel)}" not supported for model: ${modelId}, ignoring`
       )
diff --git a/apps/sim/providers/models.ts b/apps/sim/providers/models.ts
index 9c95e3b4279..cc0c7a14386 100644
--- a/apps/sim/providers/models.ts
+++ b/apps/sim/providers/models.ts
@@ -83,6 +83,12 @@ export interface ModelCapabilities {
      * to the per-provider defaults in {@link getThinkingStreamVisibility}.
      */
     streamed?: ThinkingStreamVisibility
+    /**
+     * How the pickers' `none` level reaches the API when the model rejects
+     * `thinking.type: "disabled"`. `between_tools` (Claude Sonnet 5.5) turns off
+     * up-front thinking; omitted, `none` sends no thinking config.
+     */
+    noneMode?: 'between_tools'
   }
   /** Uses native state and questions instead of a conversational prompt. */
   evaluation?: boolean
@@ -1015,7 +1021,7 @@ export const PROVIDER_DEFINITIONS: Record<string, ProviderDefinition> = {
     fileAttachment: { maxBytes: 50 * 1024 * 1024, strategy: 'remote-url' },
     name: 'Anthropic',
     description: "Anthropic's Claude models",
-    defaultModel: 'claude-sonnet-5',
+    defaultModel: 'claude-sonnet-5-5',
     modelPatterns: [/^claude/],
     icon: AnthropicIcon,
     color: '#D97757',
@@ -1071,6 +1077,30 @@ export const PROVIDER_DEFINITIONS: Record<string, ProviderDefinition> = {
         releaseDate: '2026-06-09',
         sunset: { status: 'legacy' },
       },
+      {
+        id: 'claude-sonnet-5-5',
+        pricing: {
+          input: 2.0,
+          cachedInput: 0.2,
+          output: 10.0,
+          updatedAt: '2026-09-28',
+        },
+        capabilities: {
+          forcedToolUse: false,
+          nativeStructuredOutputs: true,
+          maxOutputTokens: 128000,
+          promptCaching: { minimumCacheableTokens: 512 },
+          thinking: {
+            levels: ['low', 'medium', 'high', 'xhigh', 'max'],
+            default: 'high',
+            streamed: 'summary',
+            noneMode: 'between_tools',
+          },
+        },
+        contextWindow: 1000000,
+        releaseDate: '2026-09-28',
+        recommended: true,
+      },
       {
         id: 'claude-sonnet-5',
         pricing: {
@@ -1090,7 +1120,7 @@ export const PROVIDER_DEFINITIONS: Record<string, ProviderDefinition> = {
         },
         contextWindow: 1000000,
         releaseDate: '2026-06-30',
-        recommended: true,
+        sunset: { status: 'legacy' },
       },
       {
         id: 'claude-opus-5-5',

From 1e774221d082f38e8eb9d24260c0179ca2932cf0 Mon Sep 17 00:00:00 2001
From: Waleed Latif <walif6@gmail.com>
Date: Mon, 28 Sep 2026 12:46:44 -0700
Subject: [PATCH 2/3] test(providers): cover between_tools across Anthropic
 tool-loop turns

---
 apps/sim/providers/anthropic/core.test.ts | 89 +++++++++++++++++++++--
 1 file changed, 83 insertions(+), 6 deletions(-)

diff --git a/apps/sim/providers/anthropic/core.test.ts b/apps/sim/providers/anthropic/core.test.ts
index c8f9da76f21..60717683089 100644
--- a/apps/sim/providers/anthropic/core.test.ts
+++ b/apps/sim/providers/anthropic/core.test.ts
@@ -340,12 +340,6 @@ describe('executeAnthropicProviderRequest none thinking level', () => {
     return create.mock.calls[0][0] as Anthropic.Messages.MessageCreateParams
   }
 
-  it('sends bare between_tools thinking on claude-sonnet-5-5', async () => {
-    const payload = await sendWithThinkingNone('claude-sonnet-5-5')
-    expect(payload.thinking).toEqual({ type: 'between_tools' })
-    expect(payload.output_config).toBeUndefined()
-  })
-
   it.each(['claude-sonnet-5', 'claude-opus-5-5'])(
     'sends no thinking config on %s',
     async (model) => {
@@ -796,6 +790,89 @@ describe('streaming', () => {
       }
     })
   })
+
+  /**
+   * Both tool loops rebuild the request for every turn after a tool call, so
+   * Sonnet 5.5's `none` mapping must survive past the first request: a later
+   * turn without it would silently run adaptive thinking at the API default.
+   */
+  describe('executeAnthropicProviderRequest none thinking level across tool turns', () => {
+    const lookupTool = {
+      id: 'lookup',
+      name: 'lookup',
+      description: 'Lookup',
+      params: {},
+      parameters: { type: 'object', properties: {}, required: [] },
+    }
+    const toolTurn = message(
+      [{ type: 'tool_use', id: 'tool-1', name: 'lookup', input: {} }],
+      'tool_use'
+    )
+    const answerTurn = message([{ type: 'text', text: 'done' }], 'end_turn')
+
+    function expectBareBetweenTools(payloads: Anthropic.Messages.MessageCreateParams[]) {
+      expect(payloads).toHaveLength(2)
+      for (const payload of payloads) {
+        expect(payload.thinking).toEqual({ type: 'between_tools' })
+        expect(payload.output_config).toBeUndefined()
+      }
+    }
+
+    it('sends bare between_tools on every non-streaming turn', async () => {
+      mockExecuteTool.mockResolvedValue({ success: true, output: { value: 'tool result' } })
+      const create = vi.fn().mockResolvedValueOnce(toolTurn).mockResolvedValueOnce(answerTurn)
+
+      await executeAnthropicProviderRequest(
+        {
+          model: 'claude-sonnet-5-5',
+          apiKey: 'test-key',
+          maxTokens: 1024,
+          thinkingLevel: 'none',
+          agentEvents: true,
+          messages: [{ role: 'user', content: 'Look this up' }],
+          tools: [lookupTool],
+        },
+        {
+          providerId: 'anthropic',
+          providerLabel: 'Anthropic',
+          createClient: () => ({ messages: { create } }) as never,
+          logger: { info: vi.fn(), warn: vi.fn(), error: vi.fn(), debug: vi.fn() },
+        }
+      )
+
+      expectBareBetweenTools(create.mock.calls.map((call) => call[0]))
+    })
+
+    it('sends bare between_tools on every streaming turn', async () => {
+      mockExecuteTool.mockResolvedValue({ success: true, output: { value: 'tool result' } })
+      const createStream = vi
+        .fn()
+        .mockReturnValueOnce(stream([{ type: 'message_stop' }], toolTurn))
+        .mockReturnValueOnce(stream([{ type: 'message_stop' }], answerTurn))
+
+      const result = (await executeAnthropicProviderRequest(
+        {
+          model: 'claude-sonnet-5-5',
+          apiKey: 'test-key',
+          stream: true,
+          maxTokens: 1024,
+          thinkingLevel: 'none',
+          agentEvents: true,
+          messages: [{ role: 'user', content: 'Look this up' }],
+          tools: [lookupTool],
+        },
+        {
+          providerId: 'anthropic',
+          providerLabel: 'Anthropic',
+          createClient: () => ({ messages: { stream: createStream } }) as never,
+          logger: { info: vi.fn(), warn: vi.fn(), error: vi.fn(), debug: vi.fn() },
+        }
+      )) as StreamingExecution
+      await collectEvents(result)
+
+      expectBareBetweenTools(createStream.mock.calls.map((call) => call[0]))
+    })
+  })
 })
 
 /**

From d57f725feb2b6f00eaa61e7be19c4dc349ac65cf Mon Sep 17 00:00:00 2001
From: Waleed Latif <walif6@gmail.com>
Date: Mon, 28 Sep 2026 12:53:25 -0700
Subject: [PATCH 3/3] test(providers): assert between_tools on recorded request
 bodies

---
 apps/sim/providers/anthropic/core.test.ts | 145 ++++++++--------------
 1 file changed, 49 insertions(+), 96 deletions(-)

diff --git a/apps/sim/providers/anthropic/core.test.ts b/apps/sim/providers/anthropic/core.test.ts
index 60717683089..395c14e9fa1 100644
--- a/apps/sim/providers/anthropic/core.test.ts
+++ b/apps/sim/providers/anthropic/core.test.ts
@@ -304,51 +304,6 @@ describe('executeAnthropicProviderRequest forced tool use', () => {
   )
 })
 
-/**
- * Claude Sonnet 5.5 rejects `thinking.type: "disabled"` with a 400 and names `between_tools`
- * (which takes no effort or display field) as its lowest setting. Every other model keeps
- * `none` as "send no thinking config".
- */
-describe('executeAnthropicProviderRequest none thinking level', () => {
-  async function sendWithThinkingNone(model: string) {
-    const create = vi.fn().mockResolvedValue({
-      id: 'msg-none',
-      type: 'message',
-      role: 'assistant',
-      model,
-      content: [{ type: 'text', text: 'Done' }],
-      stop_reason: 'end_turn',
-      stop_sequence: null,
-      usage: { input_tokens: 2, output_tokens: 2 },
-    })
-    await executeAnthropicProviderRequest(
-      {
-        model,
-        apiKey: 'test-key',
-        maxTokens: 1024,
-        thinkingLevel: 'none',
-        agentEvents: true,
-        messages: [{ role: 'user', content: 'Hello' }],
-      },
-      {
-        providerId: 'anthropic',
-        providerLabel: 'Anthropic',
-        createClient: () => ({ messages: { create } }) as never,
-        logger: { info: vi.fn(), warn: vi.fn(), error: vi.fn(), debug: vi.fn() },
-      }
-    )
-    return create.mock.calls[0][0] as Anthropic.Messages.MessageCreateParams
-  }
-
-  it.each(['claude-sonnet-5', 'claude-opus-5-5'])(
-    'sends no thinking config on %s',
-    async (model) => {
-      const payload = await sendWithThinkingNone(model)
-      expect(payload.thinking).toBeUndefined()
-    }
-  )
-})
-
 describe('executeAnthropicProviderRequest native structured outputs', () => {
   /** The subset of the wire schema these assertions read back. */
   interface WireSchemaNode {
@@ -793,8 +748,11 @@ describe('streaming', () => {
 
   /**
    * Both tool loops rebuild the request for every turn after a tool call, so
-   * Sonnet 5.5's `none` mapping must survive past the first request: a later
-   * turn without it would silently run adaptive thinking at the API default.
+   * the `none` mapping must hold past the first request. Claude Sonnet 5.5
+   * rejects `thinking.type: "disabled"` and names `between_tools` (which takes
+   * no effort or display field) as its lowest setting; a later turn without it
+   * would silently run adaptive thinking. Every other model keeps `none` as
+   * "send no thinking config".
    */
   describe('executeAnthropicProviderRequest none thinking level across tool turns', () => {
     const lookupTool = {
@@ -804,28 +762,24 @@ describe('streaming', () => {
       params: {},
       parameters: { type: 'object', properties: {}, required: [] },
     }
-    const toolTurn = message(
-      [{ type: 'tool_use', id: 'tool-1', name: 'lookup', input: {} }],
-      'tool_use'
-    )
-    const answerTurn = message([{ type: 'text', text: 'done' }], 'end_turn')
+    const turns = [
+      message([{ type: 'tool_use', id: 'tool-1', name: 'lookup', input: {} }], 'tool_use'),
+      message([{ type: 'text', text: 'done' }], 'end_turn'),
+    ]
 
-    function expectBareBetweenTools(payloads: Anthropic.Messages.MessageCreateParams[]) {
-      expect(payloads).toHaveLength(2)
-      for (const payload of payloads) {
-        expect(payload.thinking).toEqual({ type: 'between_tools' })
-        expect(payload.output_config).toBeUndefined()
-      }
-    }
-
-    it('sends bare between_tools on every non-streaming turn', async () => {
+    /** Runs a tool exchange and returns every request body sent over the SDK boundary. */
+    async function runToolExchange(model: string, streaming: boolean) {
       mockExecuteTool.mockResolvedValue({ success: true, output: { value: 'tool result' } })
-      const create = vi.fn().mockResolvedValueOnce(toolTurn).mockResolvedValueOnce(answerTurn)
-
-      await executeAnthropicProviderRequest(
+      const sent: Anthropic.Messages.MessageCreateParams[] = []
+      const nextTurn = (payload: Anthropic.Messages.MessageCreateParams) => {
+        sent.push(payload)
+        return turns[sent.length - 1]
+      }
+      const result = await executeAnthropicProviderRequest(
         {
-          model: 'claude-sonnet-5-5',
+          model,
           apiKey: 'test-key',
+          stream: streaming,
           maxTokens: 1024,
           thinkingLevel: 'none',
           agentEvents: true,
@@ -835,43 +789,42 @@ describe('streaming', () => {
         {
           providerId: 'anthropic',
           providerLabel: 'Anthropic',
-          createClient: () => ({ messages: { create } }) as never,
+          createClient: () =>
+            ({
+              messages: streaming
+                ? {
+                    stream: (payload: never) =>
+                      stream([{ type: 'message_stop' }], nextTurn(payload)),
+                  }
+                : { create: async (payload: never) => nextTurn(payload) },
+            }) as never,
           logger: { info: vi.fn(), warn: vi.fn(), error: vi.fn(), debug: vi.fn() },
         }
       )
+      if (streaming) await collectEvents(result as StreamingExecution)
+      return sent
+    }
 
-      expectBareBetweenTools(create.mock.calls.map((call) => call[0]))
-    })
-
-    it('sends bare between_tools on every streaming turn', async () => {
-      mockExecuteTool.mockResolvedValue({ success: true, output: { value: 'tool result' } })
-      const createStream = vi
-        .fn()
-        .mockReturnValueOnce(stream([{ type: 'message_stop' }], toolTurn))
-        .mockReturnValueOnce(stream([{ type: 'message_stop' }], answerTurn))
-
-      const result = (await executeAnthropicProviderRequest(
-        {
-          model: 'claude-sonnet-5-5',
-          apiKey: 'test-key',
-          stream: true,
-          maxTokens: 1024,
-          thinkingLevel: 'none',
-          agentEvents: true,
-          messages: [{ role: 'user', content: 'Look this up' }],
-          tools: [lookupTool],
-        },
-        {
-          providerId: 'anthropic',
-          providerLabel: 'Anthropic',
-          createClient: () => ({ messages: { stream: createStream } }) as never,
-          logger: { info: vi.fn(), warn: vi.fn(), error: vi.fn(), debug: vi.fn() },
+    it.each([false, true])(
+      'sends bare between_tools on every turn (streaming: %s)',
+      async (streaming) => {
+        const sent = await runToolExchange('claude-sonnet-5-5', streaming)
+        expect(sent).toHaveLength(2)
+        for (const payload of sent) {
+          expect(payload.thinking).toEqual({ type: 'between_tools' })
+          expect(payload.output_config).toBeUndefined()
         }
-      )) as StreamingExecution
-      await collectEvents(result)
+      }
+    )
 
-      expectBareBetweenTools(createStream.mock.calls.map((call) => call[0]))
-    })
+    it.each(['claude-sonnet-5', 'claude-opus-5-5'])(
+      'sends no thinking config on %s',
+      async (model) => {
+        const sent = await runToolExchange(model, false)
+        expect(sent).toHaveLength(2)
+        for (const payload of sent) expect(payload.thinking).toBeUndefined()
+      }
+    )
   })
 })