Skip to content

Commit d4f6794

Browse files
wkoutrewkoutre
committed
chore: update model references to claude-sonnet-5
1 parent 3c3b90c commit d4f6794

27 files changed

Lines changed: 105 additions & 105 deletions

‎.github/REUSABLE_WORKFLOWS.md‎

Lines changed: 13 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -218,7 +218,7 @@ Notify Release (_notify-release.yml)
218218
219219
| Input | Required | Default | Description |
220220
| ------------------------------- | -------- | -------------------------------- | ---------------------------------------------------------------------------------- |
221-
| `model` | No | `'claude-sonnet-4-6'` | Claude model to use (Sonnet 4.6, Opus 4.8, or Haiku 4.5) |
221+
| `model` | No | `'claude-sonnet-5'` | Claude model to use (Sonnet 5, Opus 4.8, or Haiku 4.5) |
222222
| `allowed_tools` | No | (permissive defaults, see below) | YAML string specifying which tools Claude can use (file operations, bash commands) |
223223
| `custom_instructions` | No | `'Be sure to follow rules...'` | Additional instructions for Claude beyond CLAUDE.md files |
224224
| `timeout_minutes` | No | `'10'` | Maximum execution time in minutes (prevents runaway costs) |
@@ -282,7 +282,7 @@ The following settings are intentionally fixed to ensure consistent security and
282282

283283
- **Interactive AI Assistance**: Respond to `@claude` mentions anywhere in GitHub
284284
- **Multiple Trigger Points**: Works in issue comments, PR comments, review comments, and reviews
285-
- **Configurable Model**: Choose between Sonnet 4.6 (balanced), Opus 4.8 (thorough), or Haiku 4.5 (fast)
285+
- **Configurable Model**: Choose between Sonnet 5 (balanced), Opus 4.8 (thorough), or Haiku 4.5 (fast)
286286
- **Flexible Tool Permissions**: Control what Claude can do (read-only, read-write, or custom)
287287
- **Custom Instructions**: Add repository-specific guidelines and standards
288288
- **Bot Filtering**: Automatically excludes bot comments to prevent loops
@@ -410,7 +410,7 @@ jobs:
410410
secrets:
411411
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
412412
with:
413-
model: 'claude-sonnet-4-6'
413+
model: 'claude-sonnet-5'
414414
415415
claude-opus:
416416
if: |
@@ -511,7 +511,7 @@ Worried about Anthropic API costs
511511

512512
**Solutions**:
513513

514-
- Use `claude-sonnet-4-6` (default) instead of Opus for most tasks (~80% cheaper)
514+
- Use `claude-sonnet-5` (default) instead of Opus for most tasks (~80% cheaper)
515515
- Reduce `timeout_minutes` to limit execution time (default: 10)
516516
- The workflow includes concurrency control to prevent duplicate runs
517517
- Monitor usage at console.anthropic.com
@@ -545,7 +545,7 @@ Claude says it doesn't have permission to perform action
545545

546546
2. **Choose the Right Model**:
547547

548-
- **Sonnet 4.6** (default): Best balance of speed, capability, and cost for 90% of use cases
548+
- **Sonnet 5** (default): Best balance of speed, capability, and cost for 90% of use cases
549549
- **Opus 4.8**: Reserve for complex architectural reviews or critical security analysis
550550
- **Haiku 4.5**: Fast and cost-effective for simple questions, quick lookups, or high-volume usage
551551

@@ -606,7 +606,7 @@ jobs:
606606
secrets:
607607
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
608608
with:
609-
model: 'claude-sonnet-4-6'
609+
model: 'claude-sonnet-5'
610610
timeout_minutes: '5'
611611

612612
claude-pr-reviews:
@@ -690,7 +690,7 @@ Tips for managing Claude API costs effectively:
690690
1. **Model Selection Impact**:
691691
692692
- Haiku 4.5: Most cost-effective for simple tasks
693-
- Sonnet 4.6: ~$3 per 1M input tokens, ~$15 per 1M output tokens (default, recommended)
693+
- Sonnet 5: ~$2 per 1M input tokens, ~$10 per 1M output tokens (introductory pricing through Aug 31, 2026; default, recommended)
694694
- Opus 4.8: ~$5 per 1M input tokens, ~$25 per 1M output tokens (reserve for complex tasks)
695695
- Typical interaction: 5K-50K tokens (mostly input)
696696
@@ -721,7 +721,7 @@ Tips for managing Claude API costs effectively:
721721
722722
- 50 @claude mentions per week
723723
- Average 20K tokens per interaction
724-
- Using Sonnet 4.6
724+
- Using Sonnet 5
725725
726726
```
727727
50 mentions × 20K tokens = 1M tokens/week
@@ -984,7 +984,7 @@ jobs:
984984
| ------------------------------- | -------- | ------------------------------------ | ----------------------------------------------------------------------------------------- |
985985
| `pr_number` | Yes | - | Pull request number to review |
986986
| `base_ref` | Yes | - | Base branch name (e.g., main, master) |
987-
| `model` | No | `'claude-sonnet-4-6'` | Claude model to use for review |
987+
| `model` | No | `'claude-sonnet-5'` | Claude model to use for review |
988988
| `max_turns` | No | `15` | Maximum conversation turns for Claude |
989989
| `custom_prompt` | No | `''` | Custom prompt text (overrides file and default). Verdict logic is automatically appended. |
990990
| `custom_prompt_path` | No | `'.claude/prompts/claude-pr-bot.md'` | Path to custom prompt file in repository |
@@ -1117,7 +1117,7 @@ jobs:
11171117
pr_number: ${{ github.event.pull_request.number }}
11181118
base_ref: ${{ github.base_ref }}
11191119
# Use Opus for PRs with 'claude-opus' label, otherwise Sonnet
1120-
model: ${{ contains(github.event.pull_request.labels.*.name, 'claude-opus') && 'claude-opus-4-8' || 'claude-sonnet-4-6' }}
1120+
model: ${{ contains(github.event.pull_request.labels.*.name, 'claude-opus') && 'claude-opus-4-8' || 'claude-sonnet-5' }}
11211121
max_turns: 20 # Allow more turns for thorough Opus reviews
11221122
timeout_minutes: 45 # Longer timeout for complex reviews
11231123
secrets:
@@ -1500,7 +1500,7 @@ Claude submitted multiple reviews instead of updating one
15001500

15011501
2. **Model Selection**:
15021502

1503-
- **Sonnet 4.6** (default): Best balance for most PRs (~80% of use cases)
1503+
- **Sonnet 5** (default): Best balance for most PRs (~80% of use cases)
15041504
- **Opus 4.8**: Reserve for critical code, security-sensitive changes, or complex architectures
15051505
- Use labels to let developers request deeper reviews when needed
15061506

@@ -1584,7 +1584,7 @@ jobs:
15841584
with:
15851585
pr_number: ${{ github.event.pull_request.number }}
15861586
base_ref: ${{ github.base_ref }}
1587-
model: 'claude-sonnet-4-6'
1587+
model: 'claude-sonnet-5'
15881588
secrets:
15891589
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
15901590

@@ -1639,7 +1639,7 @@ Example cost calculation:
16391639
16401640
- 20 PRs per week
16411641
- Average 2 review iterations per PR (initial + update)
1642-
- Using Sonnet 4.6
1642+
- Using Sonnet 5
16431643
- ~30K tokens per review
16441644
16451645
```

‎.github/workflows/CLAUDE.md‎

Lines changed: 26 additions & 26 deletions
Original file line numberDiff line numberDiff line change
@@ -110,7 +110,7 @@ If both are provided, OAuth token takes precedence. At least one authentication
110110
| Input | Required | Default | Description |
111111
| ------------------------------- | -------- | ------------------------------------------------------- | ------------------------------------------------------------------- |
112112
| `prompt` | No | `""` | Direct automation prompt (enables automation mode) |
113-
| `model` | No | `claude-sonnet-4-6` | Claude model to use |
113+
| `model` | No | `claude-sonnet-5` | Claude model to use |
114114
| `allowed_tools` | No | `""` | Comma-separated list of allowed tools |
115115
| `disallowed_tools` | No | `""` | Comma-separated list of disallowed tools |
116116
| `custom_instructions` | No | CLAUDE.md instructions | Additional system prompt instructions |
@@ -145,7 +145,7 @@ You can use the `plugin_marketplaces` and `plugins` inputs to install **addition
145145
```yaml
146146
uses: Uniswap/ai-toolkit/.github/workflows/_claude-main.yml@main
147147
with:
148-
model: 'claude-sonnet-4-6'
148+
model: 'claude-sonnet-5'
149149
custom_instructions: |
150150
Focus on code quality and security.
151151
Follow CLAUDE.md guidelines.
@@ -158,7 +158,7 @@ secrets:
158158
```yaml
159159
uses: Uniswap/ai-toolkit/.github/workflows/_claude-main.yml@main
160160
with:
161-
model: 'claude-sonnet-4-6'
161+
model: 'claude-sonnet-5'
162162
secrets:
163163
CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
164164
```
@@ -168,7 +168,7 @@ secrets:
168168
```yaml
169169
uses: Uniswap/ai-toolkit/.github/workflows/_claude-main.yml@main
170170
with:
171-
model: 'claude-sonnet-4-6'
171+
model: 'claude-sonnet-5'
172172
secrets:
173173
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
174174
CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
@@ -302,7 +302,7 @@ You must enable GitHub Actions to create and approve pull requests:
302302
| `pr_number` | Yes | - | Pull request number to review |
303303
| `base_ref` | No | - | Base branch name (e.g., main, master). If not provided, fetched via GitHub API. |
304304
| `force_review` | No | `false` | Force a full review even if the code hasn't changed (bypasses patch-ID cache) |
305-
| `model` | No | `claude-sonnet-4-6` | Claude model to use for review |
305+
| `model` | No | `claude-sonnet-5` | Claude model to use for review |
306306
| `max_turns` | No | unlimited | Maximum conversation turns for Claude |
307307
| `prompt_override_files_to_skip` | No | `""` | Path to markdown file overriding "Files to Skip" section |
308308
| `prompt_override_review_priorities` | No | `""` | Path to markdown file overriding "Review Priorities" section |
@@ -387,7 +387,7 @@ uses: Uniswap/ai-toolkit/.github/workflows/_claude-code-review.yml@main
387387
with:
388388
pr_number: ${{ github.event.pull_request.number }}
389389
base_ref: ${{ github.base_ref }}
390-
model: 'claude-sonnet-4-6'
390+
model: 'claude-sonnet-5'
391391
toolkit_ref: 'main' # or 'next' to test unreleased changes
392392
secrets:
393393
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
@@ -401,7 +401,7 @@ uses: Uniswap/ai-toolkit/.github/workflows/_claude-code-review.yml@main
401401
with:
402402
pr_number: ${{ github.event.pull_request.number }}
403403
base_ref: ${{ github.base_ref }}
404-
model: 'claude-sonnet-4-6'
404+
model: 'claude-sonnet-5'
405405
secrets:
406406
CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
407407
WORKFLOW_PAT: ${{ secrets.WORKFLOW_PAT }}
@@ -560,7 +560,7 @@ If both are provided, OAuth token takes precedence. At least one authentication
560560
| `auto_commit` | No | `false` | Auto-commit and push suggestions to PR branch (bypasses suggestion_mode) |
561561
| `fail_on_missing_docs` | No | `true` | Whether missing documentation should cause workflow to fail |
562562
| `fail_on_missing_version` | No | `true` | Whether missing plugin version bumps should cause workflow to fail |
563-
| `model` | No | `claude-sonnet-4-6` | Claude model to use |
563+
| `model` | No | `claude-sonnet-5` | Claude model to use |
564564
| `max_turns` | No | unlimited | Maximum conversation turns for Claude |
565565
| `timeout_minutes` | No | `15` | Job timeout in minutes |
566566
| `toolkit_ref` | No | `main` | Git ref of ai-toolkit to use for scripts |
@@ -1067,15 +1067,15 @@ If both are provided, OAuth token takes precedence. At least one authentication
10671067

10681068
**Configuration:**
10691069

1070-
| Input | Default | Description |
1071-
| ------------------------- | ------------------- | ----------------------------------------------- |
1072-
| `branch_name` | required | Branch to work on |
1073-
| `target_branch` | `main` | Base branch for PR |
1074-
| `dry_run` | `false` | Analyze only, skip PR creation |
1075-
| `model` | `claude-sonnet-4-6` | Claude model to use |
1076-
| `timeout_minutes` | `30` | Maximum execution time |
1077-
| `debug_mode` | `true` | Show full Claude output |
1078-
| `install_uniswap_plugins` | `true` | Auto-install uniswap plugins (false to opt out) |
1070+
| Input | Default | Description |
1071+
| ------------------------- | ----------------- | ----------------------------------------------- |
1072+
| `branch_name` | required | Branch to work on |
1073+
| `target_branch` | `main` | Base branch for PR |
1074+
| `dry_run` | `false` | Analyze only, skip PR creation |
1075+
| `model` | `claude-sonnet-5` | Claude model to use |
1076+
| `timeout_minutes` | `30` | Maximum execution time |
1077+
| `debug_mode` | `true` | Show full Claude output |
1078+
| `install_uniswap_plugins` | `true` | Auto-install uniswap plugins (false to opt out) |
10791079

10801080
**Example Transformation:**
10811081

@@ -1216,13 +1216,13 @@ The Notion integration needs access to:
12161216

12171217
**Configuration Inputs:**
12181218

1219-
| Input | Default | Description |
1220-
| ------------------------ | ------------------- | ------------------------------------------------------------------------------------------------- |
1221-
| `days_back` | `7` | Number of days to look back for content |
1222-
| `model` | `claude-sonnet-4-6` | Claude model to use |
1223-
| `dry_run` | `false` | Generate but don't publish to Notion |
1224-
| `debug_mode` | `true` | Enable full Claude output for debugging |
1225-
| `slack_post_channel_ids` | `C091XE1DNP2` | Comma-separated Slack channel IDs to post newsletter announcement (e.g., C091XE1DNP2,C094URH6C13) |
1219+
| Input | Default | Description |
1220+
| ------------------------ | ----------------- | ------------------------------------------------------------------------------------------------- |
1221+
| `days_back` | `7` | Number of days to look back for content |
1222+
| `model` | `claude-sonnet-5` | Claude model to use |
1223+
| `dry_run` | `false` | Generate but don't publish to Notion |
1224+
| `debug_mode` | `true` | Enable full Claude output for debugging |
1225+
| `slack_post_channel_ids` | `C091XE1DNP2` | Comma-separated Slack channel IDs to post newsletter announcement (e.g., C091XE1DNP2,C094URH6C13) |
12261226

12271227
**Manual Trigger:**
12281228

@@ -1363,7 +1363,7 @@ jobs:
13631363
call-claude:
13641364
uses: ./.github/workflows/_claude-main.yml
13651365
with:
1366-
model: 'claude-sonnet-4-6'
1366+
model: 'claude-sonnet-5'
13671367
allowed_tools: 'read-write'
13681368
secrets:
13691369
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
@@ -1709,7 +1709,7 @@ Many workflows support manual triggering via GitHub UI or CLI:
17091709
```bash
17101710
gh workflow run claude-code-review.yml \
17111711
-f pr_number=123 \
1712-
-f model="claude-sonnet-4-6"
1712+
-f model="claude-sonnet-5"
17131713
```
17141714

17151715
## Troubleshooting

‎.github/workflows/README.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -75,7 +75,7 @@ Workflows designed to be called by other workflows using `workflow_call`. These
7575

7676
- Interactive AI assistance via @claude mentions
7777
- Works in issue comments, PR comments, review comments, and reviews
78-
- Configurable models (Sonnet 4.6, Opus 4.8, Haiku 4.5)
78+
- Configurable models (Sonnet 5, Opus 4.8, Haiku 4.5)
7979
- Flexible tool permissions (read-only, read-write, or custom)
8080
- Custom instructions for repository-specific standards
8181
- Fixed security settings (Bullfrog scanning, immutable permissions)

‎.github/workflows/_claude-code-review.yml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -70,7 +70,7 @@ on:
7070
description: "Claude model to use for review"
7171
required: false
7272
type: string
73-
default: "claude-sonnet-4-6"
73+
default: "claude-sonnet-5"
7474

7575
max_turns:
7676
description: "Maximum conversation turns for Claude. Omit to use unlimited turns."

‎.github/workflows/_claude-docs-check.yml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -83,7 +83,7 @@ on:
8383
description: "Claude model to use"
8484
required: false
8585
type: string
86-
default: "claude-sonnet-4-6"
86+
default: "claude-sonnet-5"
8787

8888
max_turns:
8989
description: "Maximum conversation turns for Claude"

0 commit comments

Comments
 (0)