FrontSources and Feeds
Source links behind our reporting, independent benchmark snapshots and API pricing notes. Source dates and availability are preserved rather than presented as newly checked on every build.
Live Wire feeds
Feed snapshot generated 2026-11-02T09:37:45Z. Status describes that collector snapshot; no public feed and a failed fetch are different conditions.
- OpenAI News: ok, 10 retained items.
- Google Blog - AI: ok, 7 retained items.
- Google DeepMind Blog: unreachable, 1 retained items.
- Meta AI Blog: no-feed, 0 retained items.
- TechCrunch AI: ok, 30 retained items.
- The Verge - AI: ok, 30 retained items.
- Ars Technica - Tech Lab: ok, 13 retained items.
- Hugging Face Blog: ok, 8 retained items.
- Mistral News: ok, 1 retained items.
- Anthropic News: ok, 1 retained items.
- xAI Blog: no-feed, 0 retained items.
- arXiv cs.CL: ok, 3 retained items.
- Simon Willison's Weblog: ok, 13 retained items.
- MarkTechPost: ok, 22 retained items.
- Google Research Blog: ok, 1 retained items.
- NVIDIA Developer Blog: ok, 10 retained items.
Independent benchmark sources
- Artificial Analysis Intelligence Index v5 (checked 1 Nov 2026): One v5 snapshot, best listed effort variant per model; effort and fallback settings are labeled. v5 launched 27 Oct and is not directly comparable with the older v4.3 scale. Older Arena and Terminal-Bench metrics were not refreshed; price/spec dates remain source-specific.
- Artificial Analysis v5 methodology announcement: Versioned methodology: nine evaluations, with a revised mix and more private test data. Do not infer model regressions from a change of index version.
- LMArena text (via AI Blogly leaderboard, 26 Oct 2026 refresh): LMArena text Elo, top ten only; models below the top ten keep their last recorded value.
- Terminal-Bench 4.0 (via Techietory snapshot): Terminal-Bench figures not on the AA leaderboard; not refreshed in this update.
- LMArena Agent arena (via BenchLeader): LMArena Agent task scores (GPT-6 Astra 14, GPT-6.1 Sol 13.6); not refreshed in this update.
Pricing sources
- Anthropic pricing Last checked:
- Anthropic prompt caching Last checked:
- OpenAI API pricing Last checked:
- OpenAI prompt caching Last checked:
- Google Gemini API pricing Last checked:
- Google Gemini context caching Last checked:
- Google Carbon announcement and pricing Last checked:
- xAI API pricing Last checked:
- xAI prompt-cache billing Last checked:
- DeepSeek API pricing and model version Last checked:
- DeepSeek automatic context caching Last checked:
- Mistral inference pricing Last checked:
- Mistral Large 4 model card Last checked:
- Mistral prompt caching Last checked:
- Kimi API pricing Last checked:
- Kimi context-cache billing Last checked:
- Alibaba Qwen API pricing Last checked:
- Qwen3.8-Max model and context limits Last checked:
- Alibaba context-cache pricing Last checked:
- Meta Muse Wonder models and API availability Last checked:
- OpenRouter model catalog (third-party; not used for calculator rates) Last checked: not verified (third-party)
Individual stories carry their own source and correction notes. Browse the reporting · Read our editorial policy.