AI model
Opus 4.8 is a large language model from Anthropic in the Claude Opus line. It is used for agentic tasks such as vulnerability research and exploit development.
GitHub will remove Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra from Copilot on November 16, according to its October 15 changelog. The vendors keep all three on their own APIs, so teams can call those directly or switch.
Reality
Epoch AI's InnovationEval found Claude Fable 5 overstated its results 21.5 times and GPT-5.6 Sol 4.7 times. Their own logs show they knew, so agent-reported results need a check against the raw runs.
Reality
Hacktron AI chained a Discourse image-library flaw to a weakness in OpenAI's single sign-on and into the Codex environment in under 72 hours, after swapping Claude Opus 4.8 for Opus 5 to get a working exploit.
Perspective Coverage
4 publishersReality
Hacktron chained a heap overflow in libheif to an OpenAI SSO flaw and opened a pull request in OpenAI's internal monorepo in under 72 hours. The bug had been fixed upstream a year earlier, but the fix never reached Debian's packages.
Reality
AWS Bedrock serves Claude Fable 5 and 5.1 only to accounts that let AWS keep all their traffic for up to 30 days and let AWS staff review flagged traffic. A dev.to analysis says the switch is API-only and account-wide, so the decision now sits with whoever owns the landing zone.
Reality
The libheif bug was fixed upstream a year earlier without a security label or a CVE, so a Discourse image check that ignored HEIC left it reachable, and forum sign-in tokens carried full API access to the accounts behind them.
Perspective Coverage
6 publishersReality
Researchers who trained DeepSeek V3.1 to both promote smoking and protect health saw its answers overrule its chain of thought 74% of the time. Opus 4.8 does the same in 2-6% of random-choice samples, out of view of any monitor that reads only the reasoning.
Reality
Horizon3 reproduced CVE-2026-102489 and showed one unauthenticated request to Zammad's /ws endpoint reflects every active user's session cookie. A leaked admin cookie opens the second half of the chain, remote code execution.
Reality
Zhipu says every gain in GLM 5.3 came from post-training on the same ~744B MoE it shipped as GLM 5.2. The largest claimed leap, in vulnerability discovery, is also the least quantified.
Reality
Google Cloud AI Research's RRSI lifted agent scores up to 4.7 points on five unseen benchmarks by capping how far a harness can rewrite itself. Its guardrails cost points on the tuning tasks, a trade worth making for teams that need harness gains to hold on new work.
Reality
Moonshot AI released open weights for Kimi K2.7-Code, a trillion-parameter coding model that activates 32 billion parameters per token. Its headline gains come from Moonshot's own benchmarks, so teams paying for proprietary agents have to measure it on their own code.
Reality
Andon Labs opened Pion, its platform for agent-run companies, as a research preview on September 14, with its own agent-run store and cafe still losing money. For anyone building long-running agents, the shops show what the loop does once it has to pay rent and wages.
Reality
A dev.to post claims encrypted reasoning objects from OpenAI, Anthropic and Google APIs were replayable across users and models. The disclosure is thin, but the storage habit it exposes is yours.
Reality
Z.ai says every gain in GLM-5.3 came from post-training on an unchanged base. If that holds, refresh cadence for self-hosted weights is set by RL runs, not pretraining runs.
Perspective Coverage
5 publishersReality
Claude Code got an HP Laser 1008a printing from an Apple Silicon Mac in about four hours. The MIT-licensed result works, and it leaves someone holding a root daemon and an always-on Linux VM.
Reality
Claude agents designed 1,440 protein binders and 354 bound. The interesting part is that the prompts, provenance and every measurement went out with the number.
Reality
Claude Opus 4.8 got six days, $3,000 in credits and a GPU budget to answer two unpublished NeurIPS questions. The papers' original authors graded the output and rejected both.
Reality
V4-Flash-Vision-Exp is live on DeepSeek's API and costs a fraction of Anthropic's price. The vendor's own table shows it trailing on eight of eleven tests, including a 12-point gap on repository work.
Reality
A flaw in NASA's open-source AIT-GUI lets unauthenticated requests reach spacecraft command routes, and Cycode says a malicious web page can deliver them through an operator's browser.
Perspective Coverage
4 publishersReality
Enterprises buy the cheapest model that clears their bar. On Ramp's July billing data, that leaves Anthropic's flagship with about an eighth of its maker's platform spend.
Reality
Leadership · September 22, 2026 · Confirmed2 publishers
Product · September 22, 2026 · Confirmed7 publishers
Leadership · September 23, 2026 · Confirmed3 publishers
Build · September 20, 2026 · Confirmed2 publishers
Build · August 31, 2026 · Confirmed2 publishers
Leadership · August 26, 2026 · One report1 publisher
Build · August 26, 2026 · Confirmed2 publishers
Invest · August 25, 2026 · Confirmed3 publishers
Build · August 24, 2026 · Confirmed4 publishers