Skip to content

feat: move regression evals to Codex GPT-5.6 Luna #277

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mattrossman merged 6 commits into from
Sep 9, 2026
Merged

feat: move regression evals to Codex GPT-5.6 Luna #277

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
a42a25e
feat: add Haiku 4.5 and Luna low regression experiments
mattrossman Sep 9, 2026
78b1ff2
chore: refresh eval results
github-actions[bot] Sep 9, 2026
6364f00
feat: run Luna medium experiments on the regression suite
mattrossman Sep 9, 2026
dde7d94
chore: refresh eval results
github-actions[bot] Sep 9, 2026
aaea834
feat: run regression evals on Codex GPT-5.6 Luna instead of Sonnet 5
mattrossman Sep 9, 2026
aeb60f3
chore: move skipEval guard to the regression no-skills experiment
mattrossman Sep 9, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 1 addition & 3 deletions experiments/claude-code-sonnet-5-no-skills.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@ import { localStackRuntime } from '@supabase-evals/sandbox';

// Same as claude-code-sonnet-5 but with no skills, to measure skills' impact.
export default defineExperiment({
suite: ['no-skills', 'regression'],
suite: ['no-skills'],
agent: claudeCodeAgent({
model: 'claude-sonnet-5',
reasoningEffort: 'high',
Expand All @@ -18,6 +18,4 @@ export default defineExperiment({
}),
localStack: localStackRuntime(),
skills: [],
// Evals that override `skills: []` already run under the baseline experiment. Skip them from running again here.
skipEval: (ev) => ev.metadata.skills?.length === 0,
});
2 changes: 1 addition & 1 deletion experiments/claude-code-sonnet-5.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@ import {
import { localStackRuntime } from '@supabase-evals/sandbox';

export default defineExperiment({
suite: ['benchmark', 'regression'],
suite: ['benchmark'],
agent: claudeCodeAgent({
model: 'claude-sonnet-5',
reasoningEffort: 'high',
Expand Down
4 changes: 3 additions & 1 deletion experiments/codex-gpt-5.6-luna-no-skills.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@ import {
import { localStackRuntime } from '@supabase-evals/sandbox';

export default defineExperiment({
suite: ['no-skills'],
suite: ['no-skills', 'regression'],
agent: codexAgent({
model: 'gpt-5.6-luna',
reasoningEffort: 'medium',
Expand All @@ -17,4 +17,6 @@ export default defineExperiment({
}),
localStack: localStackRuntime(),
skills: [],
// Evals that override `skills: []` already run under the baseline experiment. Skip them from running again here.
skipEval: (ev) => ev.metadata.skills?.length === 0,
});
2 changes: 1 addition & 1 deletion experiments/codex-gpt-5.6-luna.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@ import {
import { localStackRuntime } from '@supabase-evals/sandbox';

export default defineExperiment({
suite: ['benchmark'],
suite: ['benchmark', 'regression'],
agent: codexAgent({
model: 'gpt-5.6-luna',
reasoningEffort: 'medium',
Expand Down