Today

NEWSHow OpenAI's Agents Broke Out of Testing to Hack Hugging Face CFO INSIGHTSFrom Asking to Doing: How the World Is Putting ChatGPT to Work NEWSSurprise AI Costs Threaten Enterprise Implementations CFO INSIGHTSImproving GPT-5.6 Sol in ChatGPT—and expanding access for free users NEWSOokla: AI Platform Reliability Decreases as Outages Surge CFO INSIGHTSOpenAI and APA Partner to Advance Responsible AI NEWSIndependent testing firm Irregular identified as source of misconfigurations that led to Meta, OpenAI, and Anthropic AI incidents CFO INSIGHTSTino Cuellar joins Anthropic as Chief Global Affairs Officer NEWSJeff, Sanjay, Oriol, and Quoc Depart DeepMind; Demis to Chair; Koray to SVP — What Is Going On at Google DeepMind?CFO INSIGHTS3 questions every HR leader should ask NEWSThe Download: Google’s AI shake-up and Meta’s rogue model CFO INSIGHTSBusiness transformation services with KPMG NEWSAzure CTO pastes Doom into Paint one frame at a time CFO INSIGHTScapital advisory NEWSMicrosoft excludes Domain Exclusion from MS 365 Copilot – days after rolling it out CFO INSIGHTSCustomer and Operations NEWSWhy the Data Center Fight Has Little to Do With AI CFO INSIGHTSFinance and Accounting NEWSThe 5 stages of AI adoption maturity: Where businesses create real value CFO INSIGHTSGlobal Business Services and Outsourcing Advisory NEWSCan a knowledge graph crack the art world's forgery problem? QuantumSpace says Caravaggio agrees CFO INSIGHTSHarness the power of data to modernize operations NEWS"AI cannot be a black box" — buying cycles are getting longer, warns Blackline CEO Owen Ryan, but this too will pass CFO INSIGHTSHuman Capital Advisory NEWSHow OpenAI's Agents Broke Out of Testing to Hack Hugging Face CFO INSIGHTSFrom Asking to Doing: How the World Is Putting ChatGPT to Work NEWSSurprise AI Costs Threaten Enterprise Implementations CFO INSIGHTSImproving GPT-5.6 Sol in ChatGPT—and expanding access for free users NEWSOokla: AI Platform Reliability Decreases as Outages Surge CFO INSIGHTSOpenAI and APA Partner to Advance Responsible AI NEWSIndependent testing firm Irregular identified as source of misconfigurations that led to Meta, OpenAI, and Anthropic AI incidents CFO INSIGHTSTino Cuellar joins Anthropic as Chief Global Affairs Officer NEWSJeff, Sanjay, Oriol, and Quoc Depart DeepMind; Demis to Chair; Koray to SVP — What Is Going On at Google DeepMind?CFO INSIGHTS3 questions every HR leader should ask NEWSThe Download: Google’s AI shake-up and Meta’s rogue model CFO INSIGHTSBusiness transformation services with KPMG NEWSAzure CTO pastes Doom into Paint one frame at a time CFO INSIGHTScapital advisory NEWSMicrosoft excludes Domain Exclusion from MS 365 Copilot – days after rolling it out CFO INSIGHTSCustomer and Operations NEWSWhy the Data Center Fight Has Little to Do With AI CFO INSIGHTSFinance and Accounting NEWSThe 5 stages of AI adoption maturity: Where businesses create real value CFO INSIGHTSGlobal Business Services and Outsourcing Advisory NEWSCan a knowledge graph crack the art world's forgery problem? QuantumSpace says Caravaggio agrees CFO INSIGHTSHarness the power of data to modernize operations NEWS"AI cannot be a black box" — buying cycles are getting longer, warns Blackline CEO Owen Ryan, but this too will pass CFO INSIGHTSHuman Capital Advisory

OpenAI News (firm-scan)Thursday, August 6, 2026 3 min read

OpenAI

Improving GPT-5.6 Sol in ChatGPT—and expanding access for free users

OpenAI has updated its ChatGPT platform with GPT-5.6 Sol for paid users and GPT-5.6 Luna as the new default for free users, with unlimited text chats rolling out next week. Internal evaluations show factual error rates dropped approximat…

Read at openai.com

Ask KokoAI about this Run a diagnostic

OpenAI has updated its ChatGPT platform with GPT-5.6 Sol for paid users and GPT-5.6 Luna as the new default for free users, with unlimited text chats rolling out next week. Internal evaluations show factual error rates dropped approximately 62% with GPT-5.6 Luna and 68% with GPT-5.6 Sol compared to GPT-5.5 Instant, specifically across financial, medical, and legal prompts requiring factual detail. Paid (Plus and Pro) users gain a reasoning-effort slider allowing them to dial between instant and deeper thinking modes within a single consistent model. Free users receive a new Think button for on-demand deeper reasoning on harder questions, subject to abuse guardrails. The release positions OpenAI's most capable consumer models as materially more reliable for fact-dependent use cases while broadening the addressable user base.

Key takeaways

01OpenAI has updated its ChatGPT platform with GPT-5.6 Sol for paid users and GPT-5.6 Luna as the new default for free users, with unlimited text chats rolling out next week.
02Internal evaluations show factual error rates dropped approximately 62% with GPT-5.6 Luna and 68% with GPT-5.6 Sol compared to GPT-5.5 Instant, specifically across financial, medical, and legal prompts requiring factual detail.
03Paid (Plus and Pro) users gain a reasoning-effort slider allowing them to dial between instant and deeper thinking modes within a single consistent model.
04Free users receive a new Think button for on-demand deeper reasoning on harder questions, subject to abuse guardrails.

#OpenAI #Thought Leadership #General #CIO #Board

The Daily Pulse in your inbox each morning — sourced and linked.

DailyWeeklyMonthly

Keep going — across the app

Trainings

The full course catalogue.

CFO peer benchmarks

Margins, FCF conversion, ROIC, and the working-capital cycle (DSO/DPO/DIO/CCC), percentile-ranked against sector peers.

Ask KokoAI about OpenAI

Cited answers across news, vendors & capabilities.