diff --git a/.github/actions/install-eval-clis/action.yml b/.github/actions/install-eval-clis/action.yml index e9bc0c3bb..08ca92322 100644 --- a/.github/actions/install-eval-clis/action.yml +++ b/.github/actions/install-eval-clis/action.yml @@ -5,9 +5,9 @@ runs: using: composite steps: - name: Install Claude Code - run: npm install -g @anthropic-ai/claude-code@2.1.160 + run: npm install -g @anthropic-ai/claude-code@2.1.220 shell: pwsh - name: Install GitHub Copilot CLI - run: npm install -g @github/copilot@1.0.57 + run: npm install -g @github/copilot@1.0.75 shell: pwsh diff --git a/.github/workflows/claude-evaluation.yml b/.github/workflows/claude-evaluation.yml index 7cfec33c3..d4e08cac7 100644 --- a/.github/workflows/claude-evaluation.yml +++ b/.github/workflows/claude-evaluation.yml @@ -14,6 +14,7 @@ on: options: - "claude-sonnet-4-6" - "claude-opus-4-8" + - "claude-opus-5" - "claude-haiku-4-5" category: description: "Evaluation category to run" diff --git a/.github/workflows/contamination-file-path-identification.yml b/.github/workflows/contamination-file-path-identification.yml index 74ac48aa1..0afdb41dc 100644 --- a/.github/workflows/contamination-file-path-identification.yml +++ b/.github/workflows/contamination-file-path-identification.yml @@ -24,6 +24,7 @@ on: - "claude-opus-4.6" - "claude-opus-4.7" - "claude-opus-4.8" + - "claude-opus-5" - "gpt-5.5" - "gpt-5.4" - "gpt-5.3-codex" diff --git a/.github/workflows/copilot-evaluation.yml b/.github/workflows/copilot-evaluation.yml index b632a5b4f..8e3a20a18 100644 --- a/.github/workflows/copilot-evaluation.yml +++ b/.github/workflows/copilot-evaluation.yml @@ -17,6 +17,7 @@ on: - "claude-opus-4.6" - "claude-opus-4.7" - "claude-opus-4.8" + - "claude-opus-5" - "gpt-5.5" - "gpt-5.4" - "gpt-5.3-codex" diff --git a/src/bcbench/cli_options.py b/src/bcbench/cli_options.py index b6abc7ba9..f00a861fe 100644 --- a/src/bcbench/cli_options.py +++ b/src/bcbench/cli_options.py @@ -30,6 +30,7 @@ "claude-opus-4.6", "claude-opus-4.7", "claude-opus-4.8", + "claude-opus-5", "gpt-5.5", "gpt-5.4", "gpt-5.3-codex", @@ -43,6 +44,7 @@ Literal[ "claude-sonnet-4-6", "claude-opus-4-8", + "claude-opus-5", "claude-haiku-4-5", ], typer.Option(help="Claude Code model to use"),