Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
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 config/catalog/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -107,7 +107,7 @@ without duplicating its intrinsic identity. Virtual recipes are materialized
from packaged assets and keep their own evaluation directory.

The built-in physical inventory is curated at the creator-company level. The
current baseline contains 83 physical cards from 22 mainstream creators and
current baseline contains 84 physical cards from 22 mainstream creators and
five separately stored virtual cards. For each creator, prefer roughly the
latest three generations or representative product lines over accumulating a
shallow long tail of lesser-known creators. This policy is about Model Cards,
Expand Down
2 changes: 1 addition & 1 deletion config/catalog/catalog.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -106,9 +106,9 @@ inventory:
- nvidia/nemotron-cascade-2-30b-a3b
- publisher: OpenAI
representative_models:
- openai/gpt-6-astra
- openai/gpt-5.6-sol
- openai/gpt-5.5
- openai/gpt-5.4
- publisher: StepFun
representative_models:
- stepfun/step-3.7-flash
Expand Down
91 changes: 91 additions & 0 deletions config/catalog/resources/evaluations/single/openai.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -4912,3 +4912,94 @@
verification: claimed
source: https://cdn.openai.com/pdf/419b6906-9da6-406c-a19d-1bb078ac7637/oai_gpt-oss_model_card.pdf
redistributable: true
- id: openai/gpt-6-astra-launch-gpqa-diamond@1.0.0
model: openai/gpt-6-astra
benchmark: idavidrein/gpqa-diamond@1.0.0
benchmark_profile: published-standard
reasoning_effort: unspecified
subject:
variant: GPT-6 Astra
result_selection: maximum_across_supported_efforts
source_kind: official_launch_evaluation
metrics:
accuracy: 0.96
status: available
observed_at: 2026-09-07
evidence:
provenance: vendor_claimed
verification: claimed
source: https://openai.com/index/gpt-6-astra/
redistributable: true
- id: openai/gpt-6-astra-launch-humanitys-last-exam-tools@1.0.0
model: openai/gpt-6-astra
benchmark: cais/humanitys-last-exam@1.0.0
benchmark_profile: with-tools
reasoning_effort: unspecified
subject:
variant: GPT-6 Astra
result_selection: maximum_across_supported_efforts
tool_policy: tools_enabled
source_kind: official_launch_evaluation
metrics:
accuracy: 0.572
status: available
observed_at: 2026-09-07
evidence:
provenance: vendor_claimed
verification: claimed
source: https://openai.com/index/gpt-6-astra/
redistributable: true
- id: openai/gpt-6-astra-launch-deep-swe-1.1@1.0.0
model: openai/gpt-6-astra
benchmark: datacurve/deep-swe@1.1.0
benchmark_profile: published-agent
reasoning_effort: unspecified
subject:
variant: GPT-6 Astra
result_selection: maximum_across_supported_efforts
source_kind: official_launch_evaluation
metrics:
resolved: 0.741
status: available
observed_at: 2026-09-07
evidence:
provenance: vendor_claimed
verification: claimed
source: https://openai.com/index/gpt-6-astra/
redistributable: true
- id: openai/gpt-6-astra-launch-arc-agi-2@1.0.0
model: openai/gpt-6-astra
benchmark: arc-prize/arc-agi-2@1.0.0
benchmark_profile: semi-private
reasoning_effort: unspecified
subject:
variant: GPT-6 Astra
result_selection: maximum_across_supported_efforts
source_kind: official_launch_evaluation
metrics:
accuracy: 0.95
status: available
observed_at: 2026-09-07
evidence:
provenance: vendor_claimed
verification: claimed
source: https://openai.com/index/gpt-6-astra/
redistributable: true
- id: openai/gpt-6-astra-launch-arc-agi-1@1.0.0
model: openai/gpt-6-astra
benchmark: arc-prize/arc-agi-1@1.0.0
benchmark_profile: semi-private
reasoning_effort: unspecified
subject:
variant: GPT-6 Astra
result_selection: maximum_across_supported_efforts
source_kind: official_launch_evaluation
metrics:
accuracy: 0.985
status: available
observed_at: 2026-09-07
evidence:
provenance: vendor_claimed
verification: claimed
source: https://openai.com/index/gpt-6-astra/
redistributable: true
42 changes: 42 additions & 0 deletions config/catalog/resources/models/single/openai.yaml
Original file line number Diff line number Diff line change
@@ -1,3 +1,45 @@
- id: openai/gpt-6-astra
display_name: GPT-6 Astra
description: OpenAI flagship model for complex reasoning, coding, computer use, research, and document creation.
kind: physical
publisher: OpenAI
presentation:
logo: package:openai
monogram: O
monochrome: true
distribution:
type: proprietary_api
source: https://developers.openai.com/api/docs/models/gpt-6-astra
family: gpt-6
lifecycle: active
limits:
context_window_size: 1050000
max_output_tokens: 128000
capabilities:
- chat
- reasoning
- tools
- structured_output
- vision
- long_context
modalities:
input:
- text
- image
output:
- text
reasoning_family: gpt-6-astra
verification:
authority: OpenAI
status: claimed
verified_at: 2026-09-07
source: https://developers.openai.com/api/docs/models/gpt-6-astra
tags:
- frontier
- proprietary
- agentic
released_at: '2026-09-03'
knowledge_cutoff: '2026-04-30'
- id: openai/gpt-5.6-sol
display_name: GPT-5.6 Sol
description: OpenAI high-capability GPT-5.6 model for demanding reasoning and agentic work.
Expand Down
42 changes: 42 additions & 0 deletions config/catalog/resources/providers/openai.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,48 @@ conformance:
status: fixture_verified
verified_at: 2026-09-04
models:
- catalog: openai/gpt-6-astra
relationship: first_party
id: gpt-6-astra
protocols:
- openai/chat-completions@1
- openai/responses@1
reasoning_modes: [enabled]
reasoning_efforts: [low, medium, high, xhigh, max]
reasoning_efforts_by_protocol:
openai/chat-completions@1: [low, medium, high, xhigh]
pricing:
currency: USD
prompt_per_1m: 10.0
cached_input_per_1m: 1.0
cache_write_per_1m: 12.5
completion_per_1m: 50.0
restrictions:
tools_protocols:
- openai/responses@1
long_context_pricing:
input_threshold_tokens: 272000
prompt_multiplier: 2.0
cached_input_multiplier: 2.0
cache_write_multiplier: 2.0
completion_multiplier: 1.5
unsupported_request_fields:
openai/chat-completions@1:
- temperature
- top_p
- top_logprobs
- logprobs
openai/responses@1:
- temperature
- top_p
- top_logprobs
unsupported_include_values:
- message.output_text.logprobs
lifecycle: active
verification:
status: claimed
verified_at: 2026-09-07
source: https://developers.openai.com/api/docs/models/gpt-6-astra
- catalog: openai/gpt-5.4
relationship: first_party
id: gpt-5.4
Expand Down
6 changes: 6 additions & 0 deletions config/catalog/resources/reasoning-families.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -87,6 +87,12 @@
disabled: none
modes: [enabled, disabled]
default_mode: enabled
- id: gpt-6-astra
type: reasoning_effort
parameter: reasoning_effort
levels: [low, medium, high, xhigh, max]
modes: [enabled]
default_mode: enabled
- id: gpt-5.6
type: reasoning_effort
parameter: reasoning_effort
Expand Down
5 changes: 5 additions & 0 deletions config/catalog/schemas/catalog-resources-v1.schema.json
Original file line number Diff line number Diff line change
Expand Up @@ -112,6 +112,11 @@
"items": {"enum": ["enabled", "disabled", "adaptive"]}
},
"reasoning_efforts": {"$ref": "#/$defs/stringSet"},
"reasoning_efforts_by_protocol": {
"type": "object",
"minProperties": 1,
"additionalProperties": {"$ref": "#/$defs/stringSet"}
},
"pricing": {
"type": "object",
"additionalProperties": {"type": ["number", "string", "boolean"]}
Expand Down
Loading
Loading