diff --git a/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.test.ts b/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.test.ts
index cf0f8351d..73c6f233a 100644
--- a/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.test.ts
+++ b/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.test.ts
@@ -617,6 +617,45 @@ describe('ClaudeProvider.complete request shape', () => {
         });
     });
 
+    it('omits temperature for fable 5.1 (rejects non-default sampling)', async () => {
+        const { provider } = makeProvider();
+        messagesCreateMock.mockResolvedValueOnce(baseResponse);
+
+        await withTestActor(() =>
+            provider.complete({
+                model: 'claude-fable-5-1',
+                messages: [{ role: 'user', content: 'hi' }],
+                temperature: 0.5,
+            }),
+        );
+
+        const [args] = messagesCreateMock.mock.calls[0]!;
+        expect('temperature' in args).toBe(false);
+    });
+
+    it('forwards reasoning_effort as adaptive thinking + output_config effort on fable 5.1', async () => {
+        const { provider } = makeProvider();
+        messagesCreateMock.mockResolvedValueOnce(baseResponse);
+
+        await withTestActor(() =>
+            provider.complete({
+                model: 'claude-fable-5-1',
+                messages: [{ role: 'user', content: 'hi' }],
+                reasoning_effort: 'high',
+            } as never),
+        );
+
+        const [args] = messagesCreateMock.mock.calls[0]!;
+        // Fable 5.1 rejects `budget_tokens` and thinking is always on, so the
+        // only accepted config is adaptive; effort rides in output_config.
+        expect(args.thinking).toEqual({
+            type: 'adaptive',
+            display: 'summarized',
+        });
+        expect(args.output_config).toEqual({ effort: 'high' });
+        expect('temperature' in args).toBe(false);
+    });
+
     it('builds an enabled thinking budget from reasoning_effort on older Sonnet models', async () => {
         const { provider } = makeProvider();
         messagesCreateMock.mockResolvedValueOnce(baseResponse);
@@ -670,6 +709,22 @@ describe('ClaudeProvider model resolution', () => {
         );
     });
 
+    it('routes the bare claude-fable alias to fable 5.1 rather than fable 5', async () => {
+        const { provider } = makeProvider();
+        messagesCreateMock.mockResolvedValueOnce(baseResponse);
+
+        await withTestActor(() =>
+            provider.complete({
+                model: 'claude-fable',
+                messages: [{ role: 'user', content: 'hi' }],
+            }),
+        );
+
+        expect(messagesCreateMock.mock.calls[0]![0].model).toBe(
+            'claude-fable-5-1',
+        );
+    });
+
     it('falls back to the default model when given an unknown id', async () => {
         const { provider } = makeProvider();
         messagesCreateMock.mockResolvedValueOnce(baseResponse);
@@ -736,6 +791,46 @@ describe('ClaudeProvider.complete non-stream output', () => {
         );
     });
 
+    it('meters fable 5.1 cache reads at its reduced rate, not the 0.1x used elsewhere', async () => {
+        const { provider } = makeProvider();
+        messagesCreateMock.mockResolvedValueOnce({
+            content: [{ type: 'text', text: 'ok' }],
+            usage: {
+                input_tokens: 100,
+                output_tokens: 50,
+                cache_read_input_tokens: 1000,
+            },
+        });
+
+        await withTestActor(() =>
+            provider.complete({
+                model: 'claude-fable-5-1',
+                messages: [{ role: 'user', content: 'hi' }],
+            }),
+        );
+
+        const fable51 = CLAUDE_MODELS.find((m) => m.id === 'claude-fable-5-1')!;
+        const fable5 = CLAUDE_MODELS.find((m) => m.id === 'claude-fable-5')!;
+        // Same per-token price as Fable 5 except cache reads at a quarter of the rate.
+        expect(fable51.costs.input_tokens).toBe(fable5.costs.input_tokens);
+        expect(fable51.costs.output_tokens).toBe(fable5.costs.output_tokens);
+        expect(Number(fable51.costs.cache_read_input_tokens)).toBeCloseTo(
+            Number(fable5.costs.cache_read_input_tokens) / 4,
+        );
+
+        const [, , prefix, overrides] = recordSpy.mock.calls[0]!;
+        expect(prefix).toBe('claude:claude-fable-5-1');
+        expect(overrides.input_tokens).toBe(
+            100 * Number(fable51.costs.input_tokens),
+        );
+        expect(overrides.output_tokens).toBe(
+            50 * Number(fable51.costs.output_tokens),
+        );
+        expect(overrides.cache_read_input_tokens).toBeCloseTo(
+            1000 * Number(fable51.costs.cache_read_input_tokens),
+        );
+    });
+
     it('bills the compaction pass by summing usage.iterations', async () => {
         const { provider } = makeProvider();
         // Per Anthropic: top-level input/output reflect only the message pass;
diff --git a/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.ts b/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.ts
index e5ab3c1cd..819281e93 100644
--- a/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.ts
+++ b/src/backend/drivers/ai-chat/providers/claude/ClaudeProvider.ts
@@ -298,9 +298,10 @@ export class ClaudeProvider implements IChatProvider {
             reasoningEffort: requestedReasoningEffort,
             maxTokens: max_tokens,
         });
-        // Fable 5 and Opus 4.7/4.8 error on non-default sampling params; omit temperature entirely.
+        // Fable 5/5.1, Sonnet 5, and Opus 4.7+ reject non-default sampling; omit temperature entirely.
         // Other models require temperature=1 when thinking is enabled.
         const omitsTemperature = [
+            'claude-fable-5-1',
             'claude-fable-5',
             'claude-sonnet-5',
             'claude-opus-4-7',
@@ -313,6 +314,7 @@ export class ClaudeProvider implements IChatProvider {
               ? 1
               : (temperature ?? 0);
         const supportsEffort = [
+            'claude-fable-5-1',
             'claude-fable-5',
             'claude-sonnet-5',
             'claude-opus-5',
@@ -717,12 +719,13 @@ export class ClaudeProvider implements IChatProvider {
     }) {
         if (!reasoningEffort) return undefined;
 
-        // Fable 5, Opus 4.7/4.8, 4.6, and Sonnet 4.6 use adaptive thinking
+        // Fable 5/5.1, Opus 4.7+, 4.6, and Sonnet 4.6 use adaptive thinking
         // (`budget_tokens` is deprecated on 4.6/Sonnet 4.6, removed on
-        // Fable 5 and 4.7+). Fable 5 and Opus 4.7/4.8 omit thinking content
-        // by default; `display: 'summarized'` restores visible reasoning in
-        // the stream.
+        // Fable 5+ and Opus 4.7+). Fable 5/5.1 and Opus 4.7+ omit thinking
+        // content by default; `display: 'summarized'` restores visible
+        // reasoning in the stream.
         if (
+            modelId === 'claude-fable-5-1' ||
             modelId === 'claude-fable-5' ||
             modelId === 'claude-opus-5' ||
             modelId === 'claude-opus-4-8' ||
diff --git a/src/backend/drivers/ai-chat/providers/claude/models.ts b/src/backend/drivers/ai-chat/providers/claude/models.ts
index 06a013c9c..a54fc38cf 100644
--- a/src/backend/drivers/ai-chat/providers/claude/models.ts
+++ b/src/backend/drivers/ai-chat/providers/claude/models.ts
@@ -21,6 +21,36 @@ import type { IChatModel } from '../../types.js';
 
 // Hardcoded from https://models.dev/api.json
 export const CLAUDE_MODELS: IChatModel[] = [
+    {
+        puterId: 'anthropic:anthropic/claude-fable-5-1',
+        id: 'claude-fable-5-1',
+        modalities: { input: ['text', 'image', 'pdf'], output: ['text'] },
+        open_weights: false,
+        tool_call: true,
+        release_date: '2026-09-01',
+        aliases: [
+            'claude-fable',
+            'claude-fable-latest',
+            'claude-fable-5-1-latest',
+            'claude-fable-5.1',
+            'anthropic/claude-fable-5-1',
+        ],
+        name: 'Claude Fable 5.1',
+        costs_currency: 'usd-cents',
+        input_cost_key: 'input_tokens',
+        output_cost_key: 'output_tokens',
+        costs: {
+            tokens: 1_000_000,
+            input_tokens: 1000,
+            ephemeral_5m_input_tokens: 1000 * 1.25,
+            ephemeral_1h_input_tokens: 1000 * 2,
+            // Fable 5.1 bills cache reads at 0.025x input; every other Claude model is 0.1x.
+            cache_read_input_tokens: 1000 * 0.025,
+            output_tokens: 5000,
+        },
+        context: 1000000,
+        max_tokens: 128000,
+    },
     {
         puterId: 'anthropic:anthropic/claude-fable-5',
         id: 'claude-fable-5',
@@ -28,12 +58,7 @@ export const CLAUDE_MODELS: IChatModel[] = [
         open_weights: false,
         tool_call: true,
         release_date: '2026-06-09',
-        aliases: [
-            'claude-fable',
-            'claude-fable-latest',
-            'claude-fable-5-latest',
-            'anthropic/claude-fable-5',
-        ],
+        aliases: ['claude-fable-5-latest', 'anthropic/claude-fable-5'],
         name: 'Claude Fable 5',
         costs_currency: 'usd-cents',
         input_cost_key: 'input_tokens',