Sources
Documentation and benchmark sources behind the catalog.
Coverage in this snapshot
- Catalog models
- 168
- Ranked in the Intelligence Index
- 83
- Captured configurations, including archived results
- 291
Catalog membership means we have a source-backed model record. The homepage’s overall ranking requires a measured, non-estimated Intelligence Index result that the source has not archived. It selects the highest measured configuration for each eligible model. Different benchmarks cover different populations.
Benchmark evidence captured – . Capture dates describe the evidence, not the site publication.
Registered sources
45Alibaba official model documentationFirst Party
- Covers
- Provider, Model, Endpoint, Price
- Source ID
alibaba-docs- Policy
- #11
- Selection priority
- 90
- Reviewer
- maintainer-requested catalog expansion 2026-09-13
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
Alibaba Qwen Omni release reviewManual
- Covers
- Model, Endpoint, Price, Benchmark, Evaluation
- Source ID
alibaba-omni-release-review- Policy
- #32
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed exact model, official availability date, regional prices and hashed author benchmark tables; Realtime excluded.
Anthropic official model documentationFirst Party
- Covers
- Provider, Model, Endpoint, Price
- Source ID
anthropic-docs- Policy
- #2
- Selection priority
- 90
- Reviewer
- Codex initial adapter review
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
Anthropic Haiku 5.5 launch reportManual
- Covers
- Benchmark, Evaluation
- Source ID
anthropic-haiku-5-5-report- Policy
- #37
- Selection priority
- 90
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed exact launch table; isolated from independent AA measurements.
Anthropic Sonnet 5.5 launch reportManual
- Covers
- Benchmark, Evaluation
- Source ID
anthropic-sonnet-5-5-report- Policy
- #33
- Selection priority
- 90
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed exact launch table and conditions; isolated from independent AA results.
Artificial AnalysisThird Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
artificial-analysis- Policy
- #19
- Selection priority
- 60
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed benchmark-backed catalog expansion; official provider claims retain priority.
Reviewed author model metadataManual
- Covers
- Model
- Source ID
catalog-author-review- Policy
- #20
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Field-scoped metadata review; exact named API route; original evidence retained.
Collinear AI CWE-benchThird Party
- Covers
- Benchmark, Evaluation
- Source ID
collinear-cwe-bench- Policy
- #36
- Selection priority
- 60
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed exact models, grader/pass@k conditions and original source evidence; isolated from AA.
DeepSeek 1017 author benchmark reportFirst Party
- Covers
- Benchmark, Evaluation
- Source ID
deepseek-1017-author-benchmarks- Policy
- #46
- Selection priority
- 80
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Exact release table; author reports remain outside independent cohorts.
DeepSeek author model-card evaluationsFirst Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
deepseek-benchmarks- Policy
- #16
- Selection priority
- 60
- Reviewer
- maintainer-requested catalog expansion 2026-09-13
- Policy recorded
- Review reason
- Pinned V4 model card, exact instruct max-effort columns; no cross-source ranking
DeepSeek official model documentationFirst Party
First Party · Provider, Model, Endpoint, Price
- Covers
- Provider, Model, Endpoint, Price
- Source ID
deepseek-docs- Policy
- #12
- Selection priority
- 90
- Reviewer
- maintainer-requested catalog expansion 2026-09-13
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
First Party · Provider, Model, Endpoint, Price
- Covers
- Provider, Model, Endpoint, Price
- Source ID
deepseek-pricing- Policy
- #18
- Selection priority
- 90
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
EvalPlus authorsThird Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
evalplus- Policy
- #6
- Selection priority
- 60
- Reviewer
- Codex: apply existing reviewed benchmark registry to main catalog
- Policy recorded
- Review reason
- Reviewed pinned benchmark adapters v1; see docs/benchmarks.md
Google official model documentationFirst Party
First Party · Provider, Model, Endpoint, Price
- Covers
- Provider, Model, Endpoint, Price
- Source ID
google-argon-announcement- Policy
- #34
- Selection priority
- 90
- Reviewer
- October 2 reviewed Sol 6.1 and Argon intake
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
First Party · Provider, Model, Endpoint, Price
- Covers
- Provider, Model, Endpoint, Price
- Source ID
google-docs- Policy
- #3
- Selection priority
- 90
- Reviewer
- Codex initial adapter review
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
Google Argon benchmark reportFirst Party
- Covers
- Benchmark, Evaluation
- Source ID
google-argon-benchmark-report- Policy
- #35
- Selection priority
- 80
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed exact models, grader/pass@k conditions and original source evidence; isolated from AA.
Google published evaluationsFirst Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
google-benchmarks- Policy
- #10
- Selection priority
- 60
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed stable-model results and July 2026 methodology; docs/benchmarks.md
Google official pricingFirst Party
- Covers
- Price
- Source ID
google-pricing- Policy
- #4
- Selection priority
- 90
- Reviewer
- Codex initial adapter review
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
LiveBench authorsThird Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
livebench- Policy
- #8
- Selection priority
- 60
- Reviewer
- Reviewed LiveBench raw subtasks and exact model mappings
- Policy recorded
- Review reason
- Reviewed pinned benchmark adapters v1; see docs/benchmarks.md
Meta official model documentationFirst Party
- Covers
- Provider, Model
- Source ID
meta-docs- Policy
- #9
- Selection priority
- 90
- Reviewer
- Compare flagship author-source review
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
Xiaomi MiMo author benchmark reportsFirst Party
- Covers
- Benchmark, Evaluation
- Source ID
mimo-author-benchmarks- Policy
- #31
- Selection priority
- 80
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Reviewed missing-data import; exact identities and original evidence retained.
MMLU-Pro authorsThird Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
mmlu-pro- Policy
- #5
- Selection priority
- 60
- Reviewer
- Codex: apply existing reviewed benchmark registry to main catalog
- Policy recorded
- Review reason
- Reviewed pinned benchmark adapters v1; see docs/benchmarks.md
Moonshot AI official model documentationFirst Party
- Covers
- Provider, Model, Endpoint, Price
- Source ID
moonshot-docs- Policy
- #13
- Selection priority
- 90
- Reviewer
- maintainer-requested catalog expansion 2026-09-13
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
OpenAI official model documentationFirst Party
- Covers
- Provider, Model, Endpoint, Price
- Source ID
openai-docs- Policy
- #1
- Selection priority
- 90
- Reviewer
- Codex initial adapter review
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
OpenRouter named API offeringsFirst Party
- Covers
- Provider, Endpoint, Price
- Source ID
openrouter-offers- Policy
- #21
- Selection priority
- 90
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Field-scoped metadata review; exact named API route; original evidence retained.
Reviewed release dates: ai.google.devManual
- Covers
- Model
- Source ID
release-dates-ai-google-dev- Policy
- #24
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
Reviewed release dates: api-docs.deepseek.comManual
- Covers
- Model
- Source ID
release-dates-api-docs-deepseek-com- Policy
- #23
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
Reviewed release dates: developers.openai.comManual
- Covers
- Model
- Source ID
release-dates-developers-openai-com- Policy
- #25
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
Reviewed release dates: docs.z.aiManual
- Covers
- Model
- Source ID
release-dates-docs-z-ai- Policy
- #28
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
Reviewed release dates: openai.comManual
- Covers
- Model
- Source ID
release-dates-openai-com- Policy
- #26
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
Reviewed release dates: raw.githubusercontent.comManual
- Covers
- Model
- Source ID
release-dates-raw-githubusercontent-com- Policy
- #27
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
Reviewed release dates: www.kimi.comManual
- Covers
- Model
- Source ID
release-dates-www-kimi-com- Policy
- #30
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
Reviewed release dates: z.aiManual
- Covers
- Model
- Source ID
release-dates-z-ai- Policy
- #29
- Selection priority
- 96
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Date-only review of exact model releases; original captures and inherited field provenance preserved.
meta reviewed direct APIManual
- Covers
- Endpoint, Price
- Source ID
reviewed-direct-offers-meta- Policy
- #47
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Exact official API identity and documented introductory tariff; no currency or expiry inferred.
mistral reviewed direct APIManual
- Covers
- Endpoint, Price
- Source ID
reviewed-direct-offers-mistral- Policy
- #48
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Exact official API identity and documented introductory tariff; no currency or expiry inferred.
Reviewed endpoint identity correctionsManual
- Covers
- Endpoint
- Source ID
reviewed-endpoint-corrections- Policy
- #45
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Current official API alias identifies a different release; original observations remain immutable.
Reviewed model documentation · about.fb.comManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
reviewed-models-about-fb-com- Policy
- #39
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit October 17 release identities and exact field passages, with archived original responses.
Reviewed model documentation · api-docs.deepseek.comManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
reviewed-models-api-docs-deepseek-com- Policy
- #38
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit October 17 release identities and exact field passages, with archived original responses.
Reviewed model documentation · blogs.nvidia.comManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
reviewed-models-blogs-nvidia-com- Policy
- #41
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit October 17 release identities and exact field passages, with archived original responses.
Reviewed model documentation · docs.mistral.aiManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
reviewed-models-docs-mistral-ai- Policy
- #44
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit October 17 release identities and exact field passages, with archived original responses.
Reviewed model documentation · huggingface.coManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
reviewed-models-huggingface-co- Policy
- #43
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit October 17 release identities and exact field passages, with archived original responses.
Reviewed model documentation · reflection.aiManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
reviewed-models-reflection-ai- Policy
- #42
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit October 17 release identities and exact field passages, with archived original responses.
Reviewed model documentation · research.meta.aiManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
reviewed-models-research-meta-ai- Policy
- #40
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit October 17 release identities and exact field passages, with archived original responses.
Sierra benchmark submissionsThird Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
tau-bench- Policy
- #7
- Selection priority
- 60
- Reviewer
- Codex: apply existing reviewed benchmark registry to main catalog
- Policy recorded
- Review reason
- Reviewed pinned benchmark adapters v1; see docs/benchmarks.md
Xiaomi MiMo release reviewManual
- Covers
- Provider, Model, Endpoint, Price
- Source ID
xiaomi-release-review- Policy
- #22
- Selection priority
- 95
- Reviewer
- maintainer
- Policy recorded
- Review reason
- Explicit V2.6 API/checkpoint mappings; original official pages and pinned author metadata retained.
Z.ai author model-card evaluationsFirst Party
- Covers
- Provider, Model, Benchmark, Evaluation
- Source ID
zai-benchmarks- Policy
- #17
- Selection priority
- 60
- Reviewer
- maintainer-requested catalog expansion
- Policy recorded
- Review reason
- Pinned GLM-5.3 model card, exact GLM columns; no cross-source ranking
Z.AI official model documentationFirst Party
First Party · Provider, Model, Endpoint, Price
- Covers
- Provider, Model, Endpoint, Price
- Source ID
zai-docs- Policy
- #14
- Selection priority
- 90
- Reviewer
- maintainer-requested catalog expansion 2026-09-13
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
First Party · Provider, Model, Endpoint, Price
- Covers
- Provider, Model, Endpoint, Price
- Source ID
zai-pricing- Policy
- #15
- Selection priority
- 90
- Reviewer
- maintainer-requested catalog expansion 2026-09-13
- Policy recorded
- Review reason
- Reviewed live adapter v1; pinned allowlist, standard text inference only; see docs/live-sources.md
How we use sources
Every record keeps its original source and capture history. A source check is an observation, not certification; selection priority determines precedence, not confidence.
What missing values mean
Not reported means the reviewed source did not establish a value. It does not mean zero, free, or unsupported.
Not measured means there is no captured result for that model and setting. Estimates and archived results stay out of the homepage ranking.
Comparison limits
Benchmark versions, settings, and populations matter. Reasoning effort labels are not equal compute budgets. Ties share a rank; score gaps do not establish statistical significance.
API prices and limits apply to a specific endpoint, currency, tier, and context band. Open weights do not establish an open-source license.
Radar coverage
Model releases come from the catalog. Events and news are separately selected and summarized from primary sources, with original links and review dates in each entry. Coverage is selective, not a complete news service.
News is dated by publication or a documented story update; incident dates remain separate. Events use announced dates and times. The calendar shows events and model releases. Related models are grouped only when they share a developer announcement article.
Updates and corrections
Updates are manually captured, reviewed, and published. There is no promised refresh interval or real-time coverage. Invalid or ambiguous imports are held for review, and conflicting claims stay explicit.
A new publication can include older observations. Evidence dates change only when new information is captured or verified.
For corrections, email contact@llmcurrent.com. About and contact details.