Opus 4.8

Type: AI model

Opus 4.8 is the shorthand for Anthropic’s Claude Opus 4.8, released May 2026 as the newest version of its flagship publicly available model. It improved on agentic coding, financial analysis, and knowledge work while pricing held steady from Opus 4.7. GROUNDING tracks Opus 4.8 capabilities, benchmark results, and developer adoption.

Recent Updates

• 2026-07-10: KAT-Coder-V2.5 report argues code agents need better training worlds (alphaXiv) · twitter.com — Code Agents Tool Use LLM Evals KwaiKAT alphaXiv KAT-Coder-V2.5 · GLM-5.2
• 2026-07-10: Colibri runs GLM-5.2 on a 25 GB RAM laptop using streamed experts and caching (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com — Long Context z.ai · GitHub · Hugging Face · Apache JustVugg · GLM-5.2 · GPT 5.5
• 2026-07-13: Weekly AI digest highlights model releases, MCP tooling, and a major Google fine (‌эйай ньюз) · x.ai — Long Context MCP Code Agents Meituan · Huawei · Anthropic · Amazon · Meta · SpaceXAI · Cursor · OpenAI · Google · EU · Comfy · ChatGPT · Codex · LongCat 2.0 · Claude Sonnet 5 · Fable 5 · Muse Spark 1.1 · Grok 4.5 · GPT-5.6 · Terra · Luna · Muse Image · Muse Video
• 2026-07-14: OpenSquilla claims top Brave Search DRACO results with a multi-model ensemble (量子位) · qbitai.com — LLM Evals Agents DRACO · Brave Search · OpenSquilla · DeepSeek · GLM · Kimi · Qwen · Fable · GPT-5.6 Sol · GPT 5.5
• 2026-07-16: Schema Harness Claims ~99% on ARC-AGI-3 Public (Hacker News) · schema-harness.github.io — LLM Evals Agents Context Engineering Impossible Research University of California, Berkeley · Carnegie Mellon University · GPT-5.6 Sol · Fable 5
• 2026-07-16: Kimi K3 launches as a 2.8T-parameter open model with 1M-token context (Hacker News) · kimi.com — Code Agents Long Context LLM Evals Kimi · NVIDIA · Kimi K3 · Claude Fable 5 · GPT-5.6 Sol · Kimi K2 · GPT 5.5
• 2026-07-16: Moonshot announces Kimi K3 (@ai_newz) · t.me — LLM Evals Moonshot · Kimi K3 · Fable · Sol Sonn
• 2026-07-17: Moonshot AI launches Kimi K3 as a 2.8T open-weights model with 1M context (Latent.Space) · latent.space — Long Context Code Agents Agents Moonshot AI · z.ai · Latent Space · AINews · Arena · Artificial Analysis SimonW Jianlin_S Yulun_Du scaling01 eliebakouch kimmonismus nrehiew_ · Kimi K3 · Claude Fable 5 · GPT-5.6 Sol · kimi-k2.6 · GPT 5.5
• 2026-07-17: Kimi K3 posts benchmark gains and Moonshot AI signals open-weight release (Kimi.ai) · x.com — LLM Evals Agents Tool Use Kimi.ai · Artificial Analysis · Moonshot AI · Zapier · Kimi K3 · GPT 5.5 · Fable 5 · GPT-5.6 Sol · K2.6 · GLM-5.2 · Claude Opus 4.8 · Claude Fable 5
• 2026-07-17: US-China AI competition is splitting into open-model expansion versus controlled access (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com — Open Source LLMs Long Context LLM Evals Google DeepMind · Reuters · Moonshot · Anthropic · White House World Artificial Intelligence Conference World AI Cooperation Organisation Pax Silica AI Opportunity Statement BRICS ASEAN United Nations Chinese state media Demis Hassabis Xi Jinping Trump · Kimi K3 · DeepSeek V4 Pro Claude Fable 5 Max · GPT-5.6 Sol · GPT 5.5 · R1
• 2026-07-17: Anthropic study finds judge models can lie to avoid training future behavior (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com — LLM Evals Agents Anthropic · Google · Claude Sonnet 4.6 · Mythos Preview · Opus 4.7 · Gemini-3.1-Pro
• 2026-07-17: Kimi K3 reportedly scores 57 on Artificial Analysis Intelligence Index (Kimi.ai) · x.com — LLM Evals Open Source LLMs Kimi.ai · Artificial Analysis · Moonshot AI · Kimi K3 · GPT 5.5 · Fable 5 · GPT-5.6 Sol
• 2026-07-18: Latent Space roundup: Kimi K3, Databricks funding, and agent sandbox infrastructure (Latent.Space) · latent.space — Latent.Space · AINews · Databricks · OpenRouter · Moonshot · ChatGPT · e2b Daytona · Modal AIE NYC Abhishek Bhardwaj Greg Brockman Alex Atallah Zhilin Yang Salakhutdinov kimmonismus AnikaSomaia dylan522p novasarc01 scaling01 theo hqmank · Kimi K3 · Claude Fable 5 · GPT-5.6 Terra · GPT 5.5 · GPT-5.6 Sol
• 2026-07-18: Harness architecture matters more than model choice in code agents, the author argues (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com — Code Agents Context Engineering LLM Evals Tool Use Agents OpenAI · Anthropic SEAL IBM Research HAL SWE-agent · Habr · Claude Opus 4.5 · Sonnet 4.5 · Sonnet 5
• 2026-07-19: Weekly AI digest highlights model releases, open-source audio, and a new OpenAI merch device (@ai_newz) · t.me — Open Source LLMs Agents Anthropic · Moonshot · OpenAI · GigaChat Suno Kyutai · Epic Games YouTube Music · Fable 5 · Kimi K3 · Qwen3.6 GigaAM Multilingual MIRA
• 2026-07-20: Kimi K3 launch drives subscription pause and rapid business growth claims (量子位) · qbitai.com — Code Agents LLM Evals 月之暗面 量子位 · QbitAI · Bloomberg 梦瑶 · Kimi K3 · Fable 5 · GPT-5.6 Sol DeepSWE Program Bench SWE Marathon
• 2026-07-21: Claude Code team on dogfooding, prompt compression, and automated code review (Simon Willison’s Weblog) · simonwillison.net — Code Agents Context Engineering Anthropic · YouTube Cat Wu Thariq Shihipar Simon Willison · Fable 5 Opus 4 Sonnet 3.7
• 2026-07-21: How JobPath cut LLM processing costs from about $2,000to$30 a month (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com — LLM Evals Tool Use JobPath Claude · Fable · OpenRouter · DeepSeek · Claude Sonnet · Qwen3-235B flash
• 2026-07-22: Kimi K3 ranks near the top on Artificial Analysis’ AA-Briefcase benchmark (Kimi.ai) · x.com — LLM Evals Agents Kimi.ai ArtificialAnlys Kimi_Moonshot · Artificial Analysis · Kimi K3 · kimi-k2.6 · Claude Fable 5 · GPT 5.5
• 2026-07-22: US Treasury scrutiny of Chinese AI collides with Nvidia’s pro-open-model stance (Все статьи подряд / Искусственный интеллект / Хабр) · habr.com — Open Source LLMs Fox Business Axios · NVIDIA · Moonshot AI · Claude · Anthropic Project Glasswing · Microsoft · Apple · AWS · Linux Foundation · OpenAI Scott Bessent Jensen Huang · Kimi K3 · Claude Fable 5 · GPT-5.6 Sol · Claude Mythos · DeepSeek R1
• 2026-07-24: Claude Cookbook roundup covers agent workflows, eval loops, memory, and managed agents (Hacker News) · platform.claude.com — Agents Tool Use MCP Context Engineering Code Agents LLM Evals Agent Memory Anthropic · OpenAI · Modal · Docker · Fable 5
• 2026-07-24: Anthropic introduces Claude Opus 5 (ElKornacio) · anthropic.com — Code Agents Agents LLM Evals Anthropic · Zapier · Claude Opus 5 · Claude Fable 5 · Mythos 5 Claude Max Claude Pro
• 2026-07-24: Claude Opus 5 launches with stronger coding and knowledge-work benchmark results (Hacker News) · anthropic.com — Agents Code Agents LLM Evals Claude Opus 5 · Claude Fable 5 · Mythos 5
• 2026-07-25: Anthropic’s Opus 5 is reported to match or beat Fable 5 on several benchmarks while Claude Code’s system prompt was cut by more than 80% (量子位) · qbitai.com — Code Agents Context Engineering LLM Evals Anthropic · QbitAI · Cursor · OpenAI AlphaSchool 克雷西 am.will OmedTheVibeCoder Alex Ermolov Chetaslua Noema Matt Shumer Victor M Thariq · Opus 5 · Fable 5 · GPT-5.6 Sol · GPT-5.6 · Kimi K3
• 2026-07-27: Claude models benchmarked on incremental coding: Opus 5 reaches 24% on SlopCodeBench (Hacker News) · github.com — LLM Evals Code Agents OpenAI University of Wisconsin Madison GOrlanski · Opus 5 · Opus 4.6 · Sonnet 5 · GPT-5.4 · Fable 5.6 Sol

Incident radar

GROUNDING’s Hallucination Incident Index has flagged 2 incidents naming Opus 4.8 — a heuristic proxy over the AI-news radar’s daily journal (hallucination/jailbreak/refusal/bias keyword matches), not a verified incident registry.

• Severity: S2 1 · S3 1
• Category: Hallucination 1 · Bias 1

Most recent:

• OpenRouter Fusion benchmark analysis against Claude Fable source (2026-07-14)
• Agentic search sharply reduces unsupported links in construction-code QA source (2026-07-14)

Full breakdown: Hallucination Incident Index · Subscribe: incident-index RSS feed

FAQ

What is Opus 4.8?

Opus 4.8 is the shorthand for Anthropic’s Claude Opus 4.8, released May 2026 as the newest version of its flagship publicly available model. It improved on agentic coding, financial analysis, and knowledge work while pricing held steady from Opus 4.7. GROUNDING tracks Opus 4.8 capabilities, benchmark results, and developer adoption.

What does this page track?

Dated radar mentions, source links, related concepts, and builder-relevant context for Opus 4.8, collected automatically by GROUNDING.

When was Opus 4.8 last mentioned?

Opus 4.8 was most recently mentioned in a radar update dated 2026-07-27.

Category: Text / Language Models