GPT 5.6 Luna
PulseAugur coverage of GPT 5.6 Luna — every cluster mentioning GPT 5.6 Luna across labs, papers, and developer communities, ranked by signal.
- competes with Qwen3 8.27B 90%
- developed ChatGPT Work 90%
- instance of gpt-5.6-terra 90%
- competes with GPT-6 Luna 80%
- competes with GPT-6 Sol 80%
- competes with GPT 5.6 Terra 80%
- competes with GPT -6 Astra 70%
- other GPT -6 Astra 70%
- competes with Grok 4.5 70%
- competes with TokenPapa 70%
- competes with GLM 5.3 70%
- competes with Qwen3.8-Max 70%
- 2026-08-10 product_launch GPT 5.6 Luna has been released and is available for free on ChatGPT. source
- 2026-08-07 product_launch GPT-5.6 Luna has been released as a default model. source
- 2026-08-04 product_launch OpenAI significantly reduced the price of its GPT-5.6 Luna model by 80% and GPT-5.6 Terra by 20%. source
- 2026-08-01 product_launch OpenAI announced a 90% price reduction for its GPT-5.6 Luna model. source
- 2026-07-31 product_launch OpenAI announced an 80% price cut for its GPT-5.6 Luna model, citing internal efficiency gains. source
- 2026-07-31 product_launch OpenAI has reduced the API pricing for its GPT-5.6 Luna model. source
- 2026-07-30 product_launch OpenAI announced an 80% price reduction for its GPT-5.6 Luna model. source
- 2026-07-10 product_launch OpenAI announced the release of GPT-5.6 Luna, a new model focused on health intelligence. source
19 day(s) with sentiment data
GPT-5.6 Luna pricing cut by 80% amid Chinese LLM dominance
The recent dominance of Chinese LLMs like DeepSeek V4 Flash in global usage rankings has reportedly led OpenAI to cut GPT-5.6 Luna's price by 80%. This suggests a significant market pressure on OpenAI to remain competitive, especially as Chinese models occupy four of the top five slots on OpenRouter's weekly token usage. This observation highlights a potential shift in the global AI market power dynamics.
OpenAI to release a dedicated enterprise SKU for GPT-5.6 Luna within 60 days
Given the increasing competition from open-weight models and the reported price cuts for GPT-5.6 Luna, OpenAI may be strategically planning to release a dedicated enterprise SKU. This would allow them to differentiate their offering beyond raw performance and pricing, potentially focusing on enhanced security, compliance, or specialized tooling for enterprise clients who have lagged in adopting open models.
GPT-5.6 Luna/Sol exhibits specific failure modes in instruction following
Recent benchmarks show GPT-5.6 Luna/Sol scoring 16/18 in accuracy, with specific failure modes noted in instruction following. While Claude Opus 5 failed JSON formatting by including extra text, GPT-5.6 Luna/Sol's Python code failed due to incorrect self-tests. This indicates a need for more granular evaluation of instruction-following capabilities beyond overall accuracy, particularly in code generation and data formatting tasks.
-
Adversarial prompt increases GPT-5.6 Luna violations to 75%, HMAC authentication restores security
A study on Mastodon explored the security implications of agent handoffs in AI systems, specifically focusing on GPT-5.6 Luna. Researchers found that a single adversarial line of code could drastically increase violatio…
-
SpecStream launches live web data engine for AI agents
SpecStream has launched a new platform designed to provide AI agents with real-time access to the live web. Their Search API reportedly outperforms existing solutions on key benchmarks, achieving higher scores with smal…
-
AI agent memory benchmark sees top score, but submission faces hurdles
A team has achieved a top score on the DolphinBench benchmark for AI agent memory, outperforming the benchmark's creators, Mem0. Their agent, Empirical, scored 451 out of 600 possible points using the Hermes agent with …
-
OpenAI to retire GPT-5 and o3 models by December 2026
OpenAI has announced that several older models, including GPT-5 and o3 variants, will be shut down on December 11, 2026. This deprecation affects both base models and their fine-tuned versions, as well as specific pro m…
-
LLM code repairs can introduce silent errors, study finds
A new study reveals that Large Language Models (LLMs) can introduce silent errors when repairing code after API migrations. Even when code repairs pass typical test inputs and execute cleanly, they may produce incorrect…
-
LLM API prices mostly fall, but Google raises rates frequently · 1 source tracked
A recent analysis of LLM API pricing changes since 2024 reveals that while prices generally decrease, some models have seen significant increases, particularly in less-scrutinized pricing tiers. Google has been the most…
-
OpenAI's GPT-5.6 Luna: A Guide for Cost-Sensitive High-Volume Tasks
In November 2026, GPT-5.6 Luna is positioned by OpenAI as a cost-effective option within the GPT-5.6 family, distinct from the capability-focused Sol and balanced Terra models. This article provides a guide for develope…
-
New CompSkill method enhances LLM agent skill retrieval using execution experience
Researchers have introduced CompSkill, a novel method for improving skill retrieval in large language model (LLM) agents. Unlike traditional methods that rely solely on skill descriptions, CompSkill incorporates executi…
-
LLMs' approval decisions swayed by precedent labels, not facts
A new research paper explores how Large Language Models (LLMs) make decisions when presented with past case examples. The study found that LLMs like DeepSeek V4.1 Flash, GPT-5.6 Luna, and Claude Haiku 5.5 are significan…
-
Simon Willison releases LLM CLI tool updates with new models and AI-assisted PRs
Simon Willison has released two updates to his LLM CLI tool and Python library, version 0.37 and 0.37.1. The 0.37 release, developed over two months, introduced new models like gpt-6.1-sol and gpt-6.1-terra, improved Op…
-
OpenAI adopts mobile game-like credit system for GPT services
OpenAI has introduced a new credit system for its services, which critics liken to microtransactions found in mobile games. Users purchase credits in bundles that expire after 12 months, with different models consuming …
-
Hugging Face leaderboard shows top AI models use tools, impacting scores
A recent analysis of the Hugging Face leaderboard for "Humanity's Last Exam" (HLE) reveals that most top-performing models are utilizing external tools, a factor not explicitly tracked by the leaderboard. Out of the top…
-
New framework 'Gear' aims to simplify RLHF with scikit-learn-like interface
A new open-source framework called Gear has been developed to address challenges in Reinforcement Learning from Human Feedback (RLHF), specifically focusing on sampling, evaluation, and diversity. Inspired by meta-learn…
-
LLM cost per task varies widely, token price is misleading
A new analysis from Artificial Analysis reveals that comparing Large Language Models (LLMs) solely by their price per token is misleading. The actual cost per task can vary significantly due to factors like effort setti…
-
AI agent memory systems underperform naive retrieval, study finds · 5 sources tracked
A recent paper from UIUC, Tsinghua, and other institutions challenges the conventional wisdom that complex memory systems are superior for AI agents. The study, titled "Memory Tax: Dissecting the Net Value of Agent Memo…
-
AI security agents and SAST tools compared on new vulnerability benchmark
A new benchmark, Snyk VulnBench 2.0, evaluates the effectiveness of AI security agents and deterministic static application security testing (SAST) tools in identifying vulnerabilities. The benchmark, comprising 108 han…
-
LLM-generated stress scenarios depend on prompts, not model, study finds
A new research paper explores the use of Large Language Models (LLMs) for generating macroeconomic stress scenarios in portfolio risk simulation. The study found that LLM-generated scenarios' severity is primarily deter…
-
Anthropic files for IPO amid LLM price cuts and revenue surge
Anthropic has filed for an IPO with the SEC, revealing significant revenue growth and substantial net losses primarily due to convertible note valuations. The filing also details expenses related to a copyright settleme…
-
AI price war intensifies as open models challenge proprietary offerings
The AI market is experiencing a significant price war, with both open-weight and proprietary models reducing their costs. IBM CEO Arvind Krishna highlighted that open models now offer about 90% of the quality of top-tie…
-
Old CNN outperforms GPT-6 and Claude Fable 5.1 in skin cancer diagnosis
A 2015-era Convolutional Neural Network (CNN) has outperformed advanced large language models like GPT-6 Astra and Claude Fable 5.1 in diagnosing skin cancer from clinical and dermoscopic images. While the LLMs achieved…