# News

> Dated changelog of Signal Room data updates: model roster changes, benchmark refreshes, and fixes.

---

LLMS index: [llms.txt](/ai-market-research/llms.txt)

---

Every entry here is generated from the same changelog that drives the
[Signal Room](/ai-market-research/report/) report itself
(`data/market-state.json`) -- model roster additions, benchmark refreshes,
and fixes, in the order they happened.

---

Section pages:

- [Corrected the GPT-5.6 pricing history from OpenAI's July 30 announcement: Terra fell...](/ai-market-research/blog/2026/2026-08-07-corrected-the-gpt-5-6-pricing-history-from-openai-s-july-30/)
- [Updated DeepSeek V4-Flash to the 0731 public beta with 1M context, 384K max output...](/ai-market-research/blog/2026/2026-08-01-updated-deepseek-v4-flash-to-the-0731-public-beta-with-1m-co/)
- [Updated GPT-5.6 Terra to $2/$12 per million input/output tokens and Luna to...](/ai-market-research/blog/2026/2026-08-01-updated-gpt-5-6-terra-to-2-12-per-million-input-output-token/)
- [Added Thinking Machines Lab and its first model, Inkling: a 975B-parameter (41B...](/ai-market-research/blog/2026/2026-07-28-added-thinking-machines-lab-and-its-first-model-inkling-a-97/)
- [Added Claude Opus 5: general availability, 1M context, 128K maximum output, adaptive...](/ai-market-research/blog/2026/2026-07-25-added-claude-opus-5-general-availability-1m-context-128k-max/)
- [Refreshed current-source evidence: added OpenAI's GPT-5.6 launch table, Google's...](/ai-market-research/blog/2026/2026-07-24-refreshed-current-source-evidence-added-openai-s-gpt-5-6-lau/)
- [Added Google’s GA Gemini 3.6 Flash and Gemini 3.5 Flash-Lite with stable model IDs, 1M...](/ai-market-research/blog/2026/2026-07-22-added-google-s-ga-gemini-3-6-flash-and-gemini-3-5-flash-lite/)
- [Made the selected reference propagate through open-weight quality headlines...](/ai-market-research/blog/2026/2026-07-22-made-the-selected-reference-propagate-through-open-weight-qu/)
- [Recorded Google’s broader Flash shift: Gemini 3.5 Flash Cyber remains restricted to...](/ai-market-research/blog/2026/2026-07-22-recorded-google-s-broader-flash-shift-gemini-3-5-flash-cyber/)
- [Added Apertus-v1.1-4B-Instruct, the largest newly released Apertus Mini checkpoint...](/ai-market-research/blog/2026/2026-07-21-added-apertus-v1-1-4b-instruct-the-largest-newly-released-ap/)
- [Added Moonshot's official Kimi K3 API pricing: $3/M uncached input, $0.30/M cached...](/ai-market-research/blog/2026/2026-07-21-added-moonshot-s-official-kimi-k3-api-pricing-3-m-uncached-i/)
- [Refreshed five open agent harnesses from their canonical GitHub releases: Codex CLI...](/ai-market-research/blog/2026/2026-07-21-refreshed-five-open-agent-harnesses-from-their-canonical-git/)
- [Added Kimi K3 from Moonshot primary sources: 2.8T sparse MoE, 1M context, native...](/ai-market-research/blog/2026/2026-07-18-added-kimi-k3-from-moonshot-primary-sources-2-8t-sparse-moe/)
- [Added source-grounded GPT-5.6 Sol, Terra and Luna pricing, 1.05M context, benchmark...](/ai-market-research/blog/2026/2026-07-18-added-source-grounded-gpt-5-6-sol-terra-and-luna-pricing-1-0/)
- [Restored GPT-5.5 family visibility, verified Haiku 4.5 as the latest public Haiku...](/ai-market-research/blog/2026/2026-07-18-restored-gpt-5-5-family-visibility-verified-haiku-4-5-as-the/)
- [Updated the action queue and recommended routing: Fable for hardest retained-data...](/ai-market-research/blog/2026/2026-07-18-updated-the-action-queue-and-recommended-routing-fable-for-h/)
- [Dashboard market sweep v1.5.0: real-world re-grounding](/ai-market-research/blog/2026/2026-06-06-dashboard-market-sweep-v1-5-0-real-world-re-grounding/)
- [Restored GPT-5.3-Codex-Spark (Feb 12, 2026 release; ChatGPT Pro research preview, 128K...](/ai-market-research/blog/2026/2026-06-06-restored-gpt-5-3-codex-spark-feb-12-2026-release-chatgpt-pro/)
- [Nvidia Nemotron Coalition formed: Black Forest Labs, Cursor, LangChain, Mistral...](/ai-market-research/blog/2026/2026-06-04-nvidia-nemotron-coalition-formed-black-forest-labs-cursor-la/)
- [Nvidia releases Nemotron 3 Ultra (550B/55B MoE, hybrid Mamba-Transformer, 1M context...](/ai-market-research/blog/2026/2026-06-04-nvidia-releases-nemotron-3-ultra-550b-55b-moe-hybrid-mamba-t/)
- [Nvidia Cosmos 3 launched — open physical-AI / robotics foundation model](/ai-market-research/blog/2026/2026-06-01-nvidia-cosmos-3-launched-open-physical-ai-robotics-foundatio/)