Merge pull request #3313 from deepseek-harness/worktree-llme2e

test(e2e): use Flash for all live model coverage

packages/llm/llm-deepseek/tests/adapter.e2e.ts

@@ -28,14 +28,13 @@ import type { Config } from '@deepseek-ai/dsh-llm-deepseek'
 import { assemble, type AssembledResult } from './assemble.ts'
 
 /**
- * Real-API e2e for the direct-fetch adapter: V4 Flash + V4 Pro across
- * thinking modes and all official effort levels. The suite skips entirely
- * without $DEEPSEEK_API_KEY; the pre-release vision smoke additionally
+ * Real-API e2e for the direct-fetch adapter: V4 Flash across thinking modes
+ * and a max-effort tool round trip with reasoning passback. The suite skips
+ * entirely without $DEEPSEEK_API_KEY; the pre-release vision smoke additionally
  * requires $DEEPSEEK_VISION_E2E=1 (see vitest.e2e.config.ts).
  */
 
 const FLASH = 'deepseek-v4-flash'
-const PRO = 'deepseek-v4-pro'
 const VISION = 'deepseek-v4-flash-vision-exp'
 const VISION_E2E_ENABLED = process.env.DEEPSEEK_VISION_E2E === '1'
 const TEST_PNG = Uint8Array.from(readFileSync(
@@ -266,20 +265,23 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)('llm-deepseek e2e (real API)', ()
     expect(withThinking.usage?.reasoningTokens).toBeGreaterThan(0)
   })
 
-  it.each(['high', 'max'] as const)(
-    'pro + thinking enabled (effort %s): tool-call round trip with reasoning passback',
-    async (effort) => {
-      const ctx = await harness(PRO, { thinking: 'enabled' })
+  it(
+    'flash + thinking enabled (effort max): tool-call round trip with reasoning passback',
+    async () => {
+      const ctx = await harness(FLASH, { thinking: 'enabled' })
 
       // Turn 1: the model must call the tool (and think before it).
       const first = await assemble(ctx,{
-        model: PRO,
-        reasoningEffort: ReasoningEffortId(effort),
+        model: FLASH,
+        reasoningEffort: ReasoningEffortId('max'),
         messages: ask('What is the weather in Paris right now? Use the get_weather tool.'),
         tools: [weatherTool],
         maxTokens: 2000,
       })
-      expect(first.finish.kind).toBe('tool-calls')
+      expect(
+        first.finish.kind,
+        `DeepSeek Flash tool-call turn finished as ${JSON.stringify(first.finish)}`,
+      ).toBe('tool-calls')
       const call = first.message.content.find(block => block.type === 'tool-call')
       expect(call).toBeDefined()
       expect(call!.name).toBe('get_weather')
@@ -288,8 +290,8 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)('llm-deepseek e2e (real API)', ()
       // Turn 2: send the tool result back WITH the assistant's reasoning
       // block in history (the official thinking+tools passback rule).
       const second = await assemble(ctx,{
-        model: PRO,
-        reasoningEffort: ReasoningEffortId(effort),
+        model: FLASH,
+        reasoningEffort: ReasoningEffortId('max'),
         messages: [
           ...ask('What is the weather in Paris right now? Use the get_weather tool.'),
           createMessage({
@@ -308,22 +310,14 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)('llm-deepseek e2e (real API)', ()
         tools: [weatherTool],
         maxTokens: 2000,
       })
-      expect(second.finish.kind).toBe('stop')
+      expect(
+        second.finish.kind,
+        `DeepSeek Flash tool-result turn finished as ${JSON.stringify(second.finish)}`,
+      ).toBe('stop')
       expect(textOf(second).toLowerCase()).toMatch(/sunny|22/)
     },
   )
 
-  it('pro + thinking disabled: plain generation without reasoning blocks', async () => {
-    const ctx = await harness(PRO, { thinking: 'disabled' })
-    const result = await assemble(ctx,{
-      model: PRO,
-      messages: ask('Reply with exactly the word: pong'),
-      maxTokens: 50,
-    })
-    expect(result.finish.kind).toBe('stop')
-    expect(result.message.content.some(block => block.type === 'reasoning')).toBe(false)
-  })
-
   it('streams raw chunks in protocol order', async () => {
     const ctx = await harness(FLASH, { thinking: 'disabled' })
     const kinds: string[] = []

packages/llm/llm-deepseek/tests/adapter.spec.ts

@@ -338,7 +338,7 @@ describe('DeepSeekAdapter against a mock server', () => {
     const ctx = await harness(server.url)
 
     const result = await assemble(ctx, {
-      model: 'deepseek-v4-flash',
+      model: 'deepseek-v4-pro',
       messages: [createUserMessage({
         content: [{ type: 'text', text: 'hi' }],
         source: { kind: 'plugin', plugin: 'test' },
@@ -350,7 +350,7 @@ describe('DeepSeekAdapter against a mock server', () => {
 
     // The wire request carried the auth header contents we configured.
     expect(server.requests[0]).toMatchObject({
-      model: 'deepseek-v4-flash',
+      model: 'deepseek-v4-pro',
       max_tokens: 256_000,
       reasoning_effort: 'high',
       stream: true,

packages/llm/llm-pi-ai/tests/adapter.e2e.ts

@@ -8,14 +8,12 @@ import * as LlmDeepSeek from '@deepseek-ai/dsh-llm-deepseek'
 import { assemble, type AssembledResult } from './assemble.ts'
 
 /**
- * Real-API e2e for the pi-ai-backed adapter: V4 Flash + V4 Pro with provider
- * defaults and representative off/high/max reasoning. Mirrors the native
- * adapter's StreamChunk contract and exercises a replayed tool follow-up.
- * Key-gated.
+ * Real-API e2e for the pi-ai-backed adapter: V4 Flash defaults and
+ * off/high/max reasoning. Mirrors the native adapter's StreamChunk contract
+ * and exercises a replayed tool follow-up. Key-gated.
  */
 
 const FLASH = 'deepseek-v4-flash'
-const PRO = 'deepseek-v4-pro'
 const contexts: Context[] = []
 
 async function harness(_model: string, config: Partial<PiAiProviderProfile> = {}) {
@@ -67,10 +65,10 @@ const weatherTool: ToolSchema = {
 }
 
 describe.skipIf(!process.env.DEEPSEEK_API_KEY)('llm-pi-ai e2e (real API)', () => {
-  it.each([FLASH, PRO])('%s + provider-default reasoning: plain text generation', async (model) => {
-    const ctx = await harness(model)
+  it(`${FLASH} + provider-default reasoning: plain text generation`, async () => {
+    const ctx = await harness(FLASH)
     const result = await assemble(ctx,{
-      model,
+      model: FLASH,
       messages: ask('Reply with exactly the word: pong'),
       maxTokens: 50,
     })
@@ -91,10 +89,10 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)('llm-pi-ai e2e (real API)', () =>
     expect(textOf(result).toLowerCase()).toContain('pong')
   })
 
-  it.each([FLASH, PRO])('%s + reasoning high: reasoning blocks present', async (model) => {
-    const ctx = await harness(model)
+  it(`${FLASH} + reasoning high: reasoning blocks present`, async () => {
+    const ctx = await harness(FLASH)
     const result = await assemble(ctx,{
-      model,
+      model: FLASH,
       reasoningEffort: ReasoningEffortId('high'),
       messages: ask('Which is larger, 9.11 or 9.8? Answer with just the number.'),
       maxTokens: 2000,
@@ -104,24 +102,27 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)('llm-pi-ai e2e (real API)', () =>
     expect(textOf(result)).toContain('9.8')
   })
 
-  it('pro + reasoning max: tool-call round trip', async () => {
-    const ctx = await harness(PRO)
+  it('flash + reasoning max: tool-call round trip', async () => {
+    const ctx = await harness(FLASH)
 
     const first = await assemble(ctx,{
-      model: PRO,
+      model: FLASH,
       reasoningEffort: ReasoningEffortId('max'),
       messages: ask('What is the weather in Paris right now? Use the get_weather tool.'),
       tools: [weatherTool],
       maxTokens: 2000,
     })
-    expect(first.finish.kind).toBe('tool-calls')
+    expect(
+      first.finish.kind,
+      `pi-ai Flash tool-call turn finished as ${JSON.stringify(first.finish)}`,
+    ).toBe('tool-calls')
     const call = first.message.content.find(block => block.type === 'tool-call')
     expect(call).toBeDefined()
     expect(call!.name).toBe('get_weather')
     expect(JSON.parse(call!.arguments)).toMatchObject({ city: expect.stringMatching(/paris/i) as string })
 
     const second = await assemble(ctx,{
-      model: PRO,
+      model: FLASH,
       reasoningEffort: ReasoningEffortId('max'),
       messages: [
         ...ask('What is the weather in Paris right now? Use the get_weather tool.'),
@@ -138,7 +139,10 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)('llm-pi-ai e2e (real API)', () =>
       tools: [weatherTool],
       maxTokens: 2000,
     })
-    expect(second.finish.kind).toBe('stop')
+    expect(
+      second.finish.kind,
+      `pi-ai Flash tool-result turn finished as ${JSON.stringify(second.finish)}`,
+    ).toBe('stop')
     expect(textOf(second).toLowerCase()).toMatch(/sunny|22/)
   })
 

packages/llm/llm-pi-ai/tests/adapter.spec.ts

@@ -135,14 +135,14 @@ describe('PiAiAdapter provider routing', () => {
       thinkingBudgets: { high: 2048 },
     })
     await assemble(ctx, {
-      model: 'deepseek-v4-flash',
+      model: 'deepseek-v4-pro',
       messages: [],
       temperature: 0.2,
       maxTokens: 77,
       sessionId: 'session-for-pi' as never,
     })
     expect(server.requests[0]).toMatchObject({
-      model: 'deepseek-v4-flash',
+      model: 'deepseek-v4-pro',
       temperature: 0.2,
       max_tokens: 77,
       thinking: { type: 'enabled' },

packages/subagent/subagent-claude-code/tests/real-deepseek.e2e.ts

@@ -21,6 +21,7 @@ import * as claudeCode from '../src/index.ts'
 
 const execFileAsync = promisify(execFile)
 const OFFICIAL_DEEPSEEK_BASE_URL = 'https://api.deepseek.com'
+const DEEPSEEK_MODEL = 'deepseek-v4-flash'
 const sdkRoot = dirname(fileURLToPath(
   import.meta.resolve('@anthropic-ai/claude-agent-sdk'),
 ))
@@ -90,11 +91,11 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)(
       const env = {
         ANTHROPIC_AUTH_TOKEN: apiKey,
         ANTHROPIC_BASE_URL: `${deepSeekBaseUrl()}/anthropic`,
-        ANTHROPIC_MODEL: 'deepseek-v4-pro[1m]',
-        ANTHROPIC_DEFAULT_OPUS_MODEL: 'deepseek-v4-pro[1m]',
-        ANTHROPIC_DEFAULT_SONNET_MODEL: 'deepseek-v4-pro[1m]',
-        ANTHROPIC_DEFAULT_HAIKU_MODEL: 'deepseek-v4-flash',
-        CLAUDE_CODE_SUBAGENT_MODEL: 'deepseek-v4-flash',
+        ANTHROPIC_MODEL: DEEPSEEK_MODEL,
+        ANTHROPIC_DEFAULT_OPUS_MODEL: DEEPSEEK_MODEL,
+        ANTHROPIC_DEFAULT_SONNET_MODEL: DEEPSEEK_MODEL,
+        ANTHROPIC_DEFAULT_HAIKU_MODEL: DEEPSEEK_MODEL,
+        CLAUDE_CODE_SUBAGENT_MODEL: DEEPSEEK_MODEL,
         CLAUDE_CODE_EFFORT_LEVEL: 'max',
         CLAUDE_CONFIG_DIR: claudeConfig,
         HOME: root,

snapshots/acp/escalation-approved/cordis.yml

@@ -34,7 +34,7 @@
   name: '@deepseek-ai/dsh-acp'
   config:
     provider: deepseek-official
-    model: deepseek-v4-pro
+    model: deepseek-v4-flash
 
 - id: system-prompt
   name: '@deepseek-ai/dsh-system-prompt'