Skip to content
Open
Show file tree
Hide file tree
Changes from 1 commit
Commits
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
2 changes: 1 addition & 1 deletion prices/data.json

Large diffs are not rendered by default.

2 changes: 1 addition & 1 deletion prices/data_slim.json

Large diffs are not rendered by default.

46 changes: 43 additions & 3 deletions prices/providers/openrouter.yml
Original file line number Diff line number Diff line change
Expand Up @@ -3246,10 +3246,13 @@ models:

- id: openai/gpt-4.1-mini
match:
equals: openai/gpt-4.1-mini
or:
- equals: openai/gpt-4.1-mini
- equals: openai/gpt-4.1-mini-2025-04-14
prices:
input_mtok: 0.4
output_mtok: 1.6
cache_read_mtok: 0.1

- id: openai/gpt-4.1-nano
match:
Expand Down Expand Up @@ -3310,6 +3313,7 @@ models:
- equals: openai/gpt-5.1
- equals: openai/gpt-5.1-chat
- equals: openai/gpt-5.1-codex
- equals: openai/gpt-5-2025-08-07
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🚩 Date suffix format inconsistency across OpenRouter model aliases

The newly added date-suffixed model aliases use two different date formats: dashed dates like openai/gpt-5-2025-08-07 and openai/gpt-4.1-mini-2025-04-14, vs compact dates like openai/gpt-5.1-codex-mini-20251113, openai/gpt-5.3-codex-20260224, and openai/gpt-5.4-20260305. This appears to reflect the actual model identifiers used by OpenRouter's API rather than a formatting error, but it's worth confirming these IDs match what OpenRouter actually serves.

Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.

prices:
input_mtok: 1.25
cache_read_mtok: 0.125
Expand Down Expand Up @@ -3338,7 +3342,9 @@ models:

- id: openai/gpt-5-mini
match:
equals: openai/gpt-5-mini
or:
- equals: openai/gpt-5-mini
- equals: openai/gpt-5-mini-2025-08-07
prices:
input_mtok: 0.25
cache_read_mtok: 0.025
Expand All @@ -3364,13 +3370,47 @@ models:

- id: openai/gpt-5.1-codex-mini
match:
equals: openai/gpt-5.1-codex-mini
or:
- equals: openai/gpt-5.1-codex-mini
- equals: openai/gpt-5.1-codex-mini-20251113
prices:
input_mtok: 0.25
cache_read_mtok: 0.025
output_mtok: 2
prices_checked: 2025-11-29

- id: openai/gpt-5.3-codex
match:
or:
- equals: openai/gpt-5.3-codex
- equals: openai/gpt-5.3-codex-20260224
prices:
input_mtok: 1.75
cache_read_mtok: 0.175
output_mtok: 14

- id: openai/gpt-5.4
match:
or:
- equals: openai/gpt-5.4
- equals: openai/gpt-5.4-20260305
Comment on lines +3374 to +3399
Copy link

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The date-suffixed OpenAI model aliases use mixed date formats (e.g., openai/gpt-5-mini-2025-08-07 is YYYY-MM-DD, but openai/gpt-5.1-codex-mini-20251113 / openai/gpt-5.3-codex-20260224 / openai/gpt-5.4-20260305 are YYYYMMDD). In this file, other OpenAI date-suffixed models consistently use YYYY-MM-DD (e.g., openai/gpt-4o-2024-08-06, openai/o1-preview-2024-09-12). Please double-check the actual OpenRouter model IDs; if they are YYYY-MM-DD, update these aliases, or consider matching both formats to be robust.

Copilot uses AI. Check for mistakes.
prices:
input_mtok:
base: 2.5
tiers:
- start: 272000
price: 5
cache_read_mtok:
base: 0.25
tiers:
- start: 272000
price: 0.5
output_mtok:
base: 15
tiers:
- start: 272000
price: 22.5
Comment on lines +3404 to +3415
Copy link
Copy Markdown
Contributor

@devin-ai-integration devin-ai-integration bot Mar 29, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🚩 Tiered pricing for gpt-5.4 uses a different tier threshold than existing models

The new openai/gpt-5.4 entry uses start: 272000 for its tier thresholds, while existing tiered entries (e.g., Anthropic models like claude-sonnet-4 around lines 280-340) use start: 200000. This is likely intentional — different models can have different context window breakpoints for tiered pricing — but it's worth confirming this 272K threshold is correct for gpt-5.4.

Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.


Comment on lines +3384 to +3417
Copy link

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The newly added openai/gpt-5.3-codex and openai/gpt-5.4 entries don’t include prices_checked, while the surrounding gpt-5* models in this section do (e.g., openai/gpt-5, openai/gpt-5-mini, openai/gpt-5.1-codex-mini). Adding prices_checked for these new price definitions would keep metadata consistent and make future audits easier.

Copilot uses AI. Check for mistakes.
- id: openai/gpt-oss-120b
match:
or:
Expand Down
Loading