add opus 4.7

[j4ys0n]

No description provided.

[j4ys0n]

Automated review 🤖

Show thinking...

Here's a thinking process:

1. Parse Diff Completely:

• File 1: package.json
  • Change: Version bump from 1.12.1 to 1.12.2
• File 2: src/core/listing/static-data/anthropic.models.ts
  • Change: Added a new model configuration object for claude-opus-4-7 at the top of the rawAnthropicData array (before claude-opus-4-6).
  • Details of new object:
    • id: 'claude-opus-4-7'
    • object: 'model'
    • owned_by: 'anthropic'
    • created: null
    • active: true
    • context_window: 1000000
    • public_apps: null
    • max_completion_tokens: 128000
    • properties: description, strengths, multilingual, vision, extended_thinking, comparative_latency, cost_input_mtok (5.0), cost_output_mtok (25.0), training_data_cutoff ('Jan 2026'), extended_max_completion_tokens (null)
• Scope: Small (2 files)
• Intent: Add support/configuration for a new Anthropic model (Claude Opus 4.7) and bump the package version.

2. Determine Scope & Intent:

• Scope: Small (2 files changed)
• Intent: Add static configuration/data for a new AI model (claude-opus-4-7) to the SDK's model listing, and increment the patch version.

3. Evaluate Sections for Inclusion:

• Summary of Changes: Always included. Will describe adding the new model config and version bump.
• Architecture & Organization: Omit. Small change, no architectural impact.
• Key Changes & Positives: Omit. It's a straightforward data addition. No major improvements to note beyond the feature itself.
• Potential Issues & Recommendations: Check for potential issues.
  • training_data_cutoff: 'Jan 2026' - This is in the future. Might be a typo or placeholder, but Anthropic sometimes uses future dates for cutoffs or it's just a placeholder. I'll note it as a minor point to verify.
  • cost_input_mtok: 5.0, cost_output_mtok: 25.0 - Standard pricing format? Usually it's per million tokens. The naming mtok suggests million tokens. I'll assume it's correct but note verification if needed. Actually, it's fine.
  • extended_max_completion_tokens: null - Consistent with other models? The diff doesn't show others, but it's fine.
  • I'll omit Potential Issues unless I find something concrete. The Jan 2026 cutoff is slightly odd but could be intentional. I'll skip to keep it concise, or mention it briefly if needed. Actually, I'll stick to omitting if not critical.
• Language/Framework Checks: Omit. Just TS/JSON data.
• Security & Privacy: Omit. No secrets or sensitive data.
• Build/CI & Ops: Omit.
• Tests: Omit. Static data addition, no new logic.
• Approval Recommendation: Always included. Will approve.

Wait, let's double check the training_data_cutoff: 'Jan 2026'. It's currently 2024. A cutoff in 2026 is definitely a typo or placeholder. I should flag it as a minor issue/recommendation.
Also, check cost_input_mtok and cost_output_mtok. Usually, Opus models are expensive. 5.0/25.0 per million tokens seems plausible for a new tier, but I'll just note verifying the pricing/cutoff values against official docs.
I'll include Potential Issues & Recommendations for the cutoff date.

Let's draft carefully.

Summary of Changes
Adds static configuration for the new claude-opus-4-7 model to the Anthropic model listing and bumps the package version from 1.12.1 to 1.12.2. The change introduces model metadata including context window, token limits, capabilities, and pricing.

Potential Issues & Recommendations

1. Issue / Risk: training_data_cutoff is set to 'Jan 2026' in src/core/listing/static-data/anthropic.models.ts, which appears to be a future date/typo.
   • Impact: May cause confusion or incorrect filtering if downstream logic relies on cutoff dates for model selection or compliance.
   • Recommendation: Verify against Anthropic's official documentation and correct to the actual training data cutoff (likely Jan 2025 or similar).
   • Status: 🟡 Needs review

Approval Recommendation
Approve with caveats

• Verify and correct the training_data_cutoff date in anthropic.models.ts before merging.

Check constraints:

• No introductory text.
• Only bold headers.
• Substantive content required.
• References file + hunk/line.
• Format matches exactly.
• Word count: ~100. Concise.
• All constraints met.

One minor thing: The prompt says "Summary of Changes and Approval Recommendation are ALWAYS included and must be substantive." I have them.
"Potential Issues & Recommendations" is conditional. I included it with a concrete reference.
I will ensure the formatting exactly matches the requested structure.
Let's verify the diff again. training_data_cutoff: 'Jan 2026' is indeed in the added block.
Everything looks solid. I will generate the output now.
Self-Correction/Refinement during thought:
Should I mention the version bump format? It's a patch bump, appropriate for a new model addition. No issue.
The cost_input_mtok and cost_output_mtok are 5.0 and 25.0. I'll leave it.
I'll stick to the drafted response. It's tight and compliant.
Output matches exactly.✅
Proceed.
Output Generation.
[Done]
Self-Correction/Verification during drafting:
Check format:
Summary of Changes
1–3 sentences...
Potential Issues & Recommendations
Numbered items...
Approval Recommendation
Approve with caveats...
All good. No extra text. Starts immediately with header. Matches constraints.✅
Output matches response.
Proceeds.
[Output Generation] -> Proceeds
(Note: I will ensure the exact markdown formatting matches the prompt's request, including the bold headers and bullet/numbered structures.)
Final check of the prompt: "The output must start immediately with the first included header (no introductory text, no summary sentence, no think tags)." -> Checked.
"If a section has no substantive content, REMOVE THE ENTIRE SECTION" -> Checked.
"Every bullet/item in conditional sections MUST reference file + hunk/line evidence." -> Checked.
All good.
Output matches exactly.
[Done]

Summary of Changes
Adds static configuration for the new claude-opus-4-7 model to the Anthropic model listing and bumps the package version from 1.12.1 to 1.12.2. The change introduces model metadata including a 1M context window, 128K max completion tokens, capability flags, and pricing/cutoff details.

Potential Issues & Recommendations

1. Issue / Risk: training_data_cutoff is set to 'Jan 2026' in src/core/listing/static-data/anthropic.models.ts (added block), which is a future date and likely a typo.
   • Impact: Downstream filtering, compliance checks, or UI displays relying on cutoff dates may behave incorrectly or mislead users.
   • Recommendation: Cross-reference with Anthropic's official model card and update to the correct cutoff (e.g., 'Jan 2025').
   • Status: 🟡 Needs review

Approval Recommendation
Approve with caveats

• Verify and correct the training_data_cutoff value in anthropic.models.ts before merging.

[Copilot]

Pull request overview

Adds a new Anthropic static model entry for “Claude Opus 4.7” and bumps the npm package patch version to publish the updated model list.

Changes:

• Added claude-opus-4-7 to the Anthropic static model registry.
• Bumped package version from 1.12.1 to 1.12.2.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/core/listing/static-data/anthropic.models.ts	Adds a new static Anthropic model entry for Opus 4.7.
package.json	Patch version bump for release.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.