- Start
- May 28, 202695% CONFIDENCEfrom anthropic.com
Claude Opus 4.8 Released
- Anthropic released Claude Opus 4.8 on May 28, 2026, building on Opus 4.7 at the same $5/$25 per million token price; API name claude-opus-4-8[1][3]
- Fast mode (2.5x speed) is priced at $10/$50 per million tokens, three times cheaper than for previous models[1]
- Launched alongside dynamic workflows in Claude Code (research preview) running hundreds of parallel subagents, and effort control in claude.ai and Cowork on all plans[1]
- Anthropic said Mythos-class models were expected for all customers in the coming weeks[1]
- The Opus product page lists Opus 4.8 (May 28, 2026) and later Opus 5[3]
Notable features
- Opus 4.8 builds on Opus 4.7 with improvements across benchmarks and is described as a more effective collaborator; Anthropic calls it a modest but tangible improvement on its predecessor[1]
- Specifications: 1M-token context window, 128K max output tokens (300K on the Batch API in beta), adaptive thinking with high default effort, text and image input with text output, and a January 2026 reliable knowledge cutoff[2]
- Honesty: about four times less likely than its predecessor to let flaws in its own code pass unremarked, and misaligned behavior is substantially lower than Opus 4.7[1]
- Defaults to high effort, with "extra" (xhigh in Claude Code) or max for harder tasks, and higher Claude Code rate limits[1]
- Tester figures: Online-Mind2Web 84%; Databricks Genie 61% cheaper token cost than Opus 4.7[1]
References 390% CONFIDENCE
The first entry is always the pin's source. Overall confidence is a weighted average of how firmly each reference supports the start and end times used above; a reference counts half as much for every 180 days older than the newest.
- [1]95%Claude Opus 4.8anthropic.com· Posted Sep 19, 2026· Starts May 28, 2026· 33% of score
Anthropic's[3] own announcement, dated the day the model became available.
- [2]90%Claude Opus 4.8 - Claude Platform Docsplatform.claude.com· Added Sep 22, 2026· 34% of score
Anthropic's[1][3] Claude Opus 4.8 model page lists a 1M-token context window, 128K max output tokens, adaptive thinking with high default effort, and a Jan 2026 reliable knowledge cutoff.
- [3]85%Claude Opus \ Anthropicanthropic.com· Added Sep 19, 2026· Starts May 28, 2026· 33% of score
Anthropic's[1] Claude[2] Opus page lists this announcement and its date.
Suggest a correction
Something missing or wrong? Say it in your own words: a link that backs this pin up, a different start or end date and why, or a fact it lacks or gets wrong. The AI checks it against this pin's sources, searches for better ones, and adds any page that backs you up. The pin's own sources still count most.
