Opus 4.7 thinking summaries not rendered (VSCode extension 2.1.112)

Thinking summaries not rendering for claude-opus-4-7[1m] in VS Code extension 2.1.112, while they worked on Opus 4.6, is likely a version-specific rendering path issue — the extension may have a model version check for thinking summary rendering that was not updated when 4.7 rolled out, causing the renderer to fall through to a non-thinking branch.

From an agent capability perspective, thinking summaries are a significant quality-of-experience feature for users who use extended thinking — without them, Opus 4.7 on max effort appears identical to lower-effort modes in the UI, removing a key signal about what the model is doing.

Narrowing the failure:

1. Check extension logs: The log entry thinkingLevel:"default_on" suggests the thinking is configured but may not be rendering. Look for any warning about unknown response content types in the extension log.

1. Check if thinking is actually happening: The raw API response from Opus 4.7 should include thinking content blocks if extended thinking is active. If those blocks are present in the response but not rendered, it is a rendering bug. If they are absent, the model itself is not doing extended thinking despite the configuration.

1. Test in CLI mode: Run the same prompt via claude --model claude-opus-4-7[1m] in terminal with --output-format json. If the JSON response includes thinking blocks, the rendering bug is VS Code extension-specific. If no thinking blocks, the issue is in how the extension configures the model request.

The 4.6 → 4.7 model identifier format change (claude-opus-4-7[1m] vs previous) may be the culprit — the extension's thinking summary renderer might have a model version check that only enables the thinking UI for known-good model IDs.

dupe of #49322 (see https://github.com/anthropics/claude-code/issues/49322#issuecomment-4265369864 for workaround)

Same issue. Strongly need this feature.

Same issue here as well. Would prefer to not have to rely on a wrapper script to see thinking when I want to use it in vs code.

Hitting this too on extension 2.1.112 with claude-opus-4-7[1m]. Confirmed showThinkingSummaries: true, effortLevel: "max", viewMode: "verbose" all ignored. No chevron, no summary, even on reasoning-heavy turns. Log shows thinkingLevel: "default_on" so the setting is reaching the binary, but the model descriptor strips the capability downstream. Downgrading to claude-opus-4-6[1m] restores thinking display immediately, which matches the missing model-picker branch diagnosis in the OP. +1 for adding the 4.7 entry to the in-binary capability table.

It is not related to the missing branch the OP talks about, which is just a red herring. It's due to an intentional server-side change to disable thinking summaries by default specifically for Opus 4.7.

See the other thread #49322, which this is a dupe of, for technical details and possible workarounds.

Same bug on Windows / VS Code extension 2.1.x with Opus 4.7. Not macOS-specific.

Root cause is now public (credit to the workaround in #8477 (comment)): Opus 4.7 changed the API default for thinking blocks from display: "summarized" to display: "omitted". The blocks are still in the response stream, but the thinking field is empty unless the caller passes display: "summarized" explicitly. Documented in Anthropic's own migration guide: https://platform.claude.com/docs/en/about-claude/models/migration-guide#migrating-to-claude-opus-4-7

Net effect: showThinkingSummaries, alwaysThinkingEnabled, and viewMode: "verbose" have nothing to render, because Claude Code does not pass display: "summarized" when calling the Opus 4.7 API. showThinkingSummaries is the documented, official way to surface thinking in the VS Code extension, and it is silently broken on Opus 4.7. The settings surface never caught up to the API change.

Two undocumented workarounds (from the same comment):

1. Add to ~/.claude/settings.json:

``json
{
"env": {
"CLAUDE_CODE_EXTRA_BODY": "{\"thinking\":{\"type\":\"adaptive\",\"display\":\"summarized\"}}"
}
}
``

1. Hidden CLI flag: --thinking-display summarized

The env var approach works for the VS Code extension since it delegates to the CLI.

Why this is high-priority, not a cosmetic regression:

Visible real-time thinking was load-bearing for the steering loop. With it visible, you could stop the model going down a wrong branch before it spent N turns of token budget on a wrong premise; you could verify it had internalized constraints from CLAUDE.md / memory rather than hallucinated them; and you could audit assumptions before committing to a plan.

With it hidden behind an unexpandable pill, you cannot tell whether a wrong final answer came from a wrong premise or a wrong execution of a right premise. Silent assumption drift becomes invisible until after the wrong action is taken. "Thought for 376s" is a duration, not a signal. It tells you nothing actionable. This is a regression in the developer's ability to supervise the agent.

On the user side: this is a paid product. Pro at $20/month, Max at $100–$200/month. Paying customers should not have to monkey-patch cli.js, reverse-engineer an undocumented CLAUDE_CODE_EXTRA_BODY env var, or scrape issue threads to recover functionality that worked in the previous release of the same product. The workaround being undocumented is itself a bug. The gap between what the API supports and what Claude Code exposes as a first-class setting keeps widening with every model bump.

Related open issues, same root problem:

• #8477 — original tracking issue where the API-default change was traced and the workaround was published
• #49268 — harness doesn't set display: "summarized" for Opus 4.7 (same root cause, named explicitly)
• #49757 — empty/unexpandable stub in VS Code extension (Opus 4.7)
• #49322 — thinking summaries not rendered in VS Code extension (Opus 4.7)
• #48065 — thinking summaries not displayed when showThinkingSummaries enabled
• #49739 — toggle button inverted, no visual feedback (v2.1.112), likely a related UI-state bug
• #30958 — earlier transcript/TUI variant of the same regression (v2.1.69)
• #51131 — extended thinking dropdown no longer expandable in VS Code extension (Opus 4.7)
• #33163 — "Bring back thinking" feature request, same underlying ask

What would actually resolve this:

1. Pass display: "summarized" by default for Opus 4.7+ when showThinkingSummaries and alwaysThinkingEnabled are on. Or expose a first-class thinkingDisplay setting alongside the existing thinkingEnabled and effortLevel.
2. Restore the expandable thinking block in the VS Code extension chat panel so showThinkingSummaries actually shows summaries when the data is present.
3. Update the docs so they reflect current showThinkingSummaries semantics on Opus 4.7+ instead of describing behavior that no longer applies.

Hi team,

Reporting that issue #49902 (Opus 4.7 thinking summaries not rendered in the VSCode extension) is still present in the latest published version, 2.1.126 — not just 2.1.112 as originally reported.

Environment

OS: Windows 11 Pro 26200
Extension: anthropic.claude-code-2.1.126-win32-x64
Model in use: claude-opus-4-7[1m]
Symptom: no thinking blocks appear in the conversation panel, even with Extended Thinking toggled on. The same model in the terminal CLI renders thinking correctly. Switching to claude-opus-4-6[1m] in the same extension immediately restores thinking rendering.
Root cause (confirmed by inspecting resources/native-binary/claude.exe)

The bundled JS inside the binary contains the model-resolution switch:

else if (_ === "opus" && Rz()) return UVH([...q, Sl7(!1)]);
else if (_ === "opus[1m]" && Rz()) return UVH([...q, sVq(!1)]);
else if (_ === "claude-opus-4-6" && Rz()) return UVH([...q, El7(H,!1)]);
else if (_ === "claude-opus-4-6[1m]" && Rz()) return UVH([...q, hl7(H,!1)]);
else { / falls through to {value:_, label:_, description:"Custom model"} / }
There is no branch for "claude-opus-4-7" or "claude-opus-4-7[1m]", so 4.7 falls through into the generic Custom model path, which strips the capability metadata required to render thinking blocks.

The metadata factory function for 4.7 — Vl7() — already exists in the binary (it returns {label: "Opus 4.7 (1M context)", ...}), it is just never invoked because the corresponding switch arms are missing.

Suggested fix

Add the two missing branches before the fallthrough, mirroring the 4.6 ones:

else if (_ === "claude-opus-4-7" && Rz()) return UVH([...q, / Opus 4.7 factory, non-1M / ]);
else if (_ === "claude-opus-4-7[1m]" && Rz()) return UVH([...q, Vl7(H,!1)]);
Counts of the relevant string in the shipped 2.1.126 binary, for reference:

claude-opus-4-6[1m]: 6 occurrences
claude-opus-4-7[1m]: 0 occurrences
Happy to provide byte offsets if useful. Thanks for the great tool overall.

Three weeks, and still no fix.

This is a serious issue, and causing divergence and failures because I cant spot when Opus 4.7 is off track.

I would really appreciate a fix to this issue its my single biggest daily problem.

this is still broken on 2.1.132 and is significantly impacting my workflow.

The thinking display is critical for me because:

• I rely on visible reasoning to catch alignment issues and verify the model's chain of thought
• The workaround (downgrading to claude-opus-4-6 in settings.json) appears to also break the 1M context window — context fills up at ~200K, suggesting the model-picker bug also affects context window mapping
• Rolling back to 2.1.109 forces me to give up Opus 4.7 entirely

This issue has been open for 3+ weeks and looks like a missing branch in the model-picker logic. Could we get a status update on the fix timeline? Even a temporary toggle (e.g. "forceShowThinking": true in settings.json) would unblock affected users while a proper fix is in progress.

Affected: VS Code extension 2.1.112 through 2.1.132, Opus 4.7 model.

Seems like workaround number 1 (CLAUDE_CODE_EXTRA_BODY) breaks Web Search tool :(

{
  "env": {
    "CLAUDE_CODE_EXTRA_BODY": "{\"thinking\":{\"type\":\"adaptive\",\"display\":\"summarized\"}}"
  }
}

Getting

 400 400 {"type":"error","error":{"type":"invalid_request_error","message":"Thinking may not be enabled when tool_choice forces tool use."},"request_id":"<id>"}

@zaitsman I filed #56984 for this a couple days ago

Filed #59844 with a one-line CLI fix (preferred) or one-line extension fix (alternative) that resolves the symptom here and across the related thread (#49322 / #49268 / #51131 / #49757 / #48065 / #49739 / #33163 / #8477 / #30958).

The CLI-side fix: drop the !getIsNonInteractiveSession() gate from the K3.display = "summarized" assignment, so showThinkingSummaries: true actually does what its name promises in non-interactive surfaces (VS Code chat panel, SDK, --print).

Structurally distinct from the CLAUDE_CODE_EXTRA_BODY env-var workaround that #56984 documents breaking WebSearch / WebFetch: both fixes in #59844 set only display, not type, so the CLI's per-request thinking-disable logic continues to work for forced-tool-use and incompatible-model sub-calls. Verified locally on v2.1.142.

Disclosure: written with Claude Opus 4.7.

see https://github.com/anthropics/claude-code/issues/49322#issuecomment-4823842675

this issue (which is a dupe of #49322) is now fixed since vs code extension version 2.1.176

This bug appears again for Version 2.1.205