# Claude Opus 4.8

Claude Opus 4.8 ialah model daripada Anthropic, dikeluarkan pada 2026-05-28. Tetingkap konteks ialah 1,000,000 token. Keluaran maksimum ialah 128,000 token. Harganya ialah $5.00 bagi sejuta token input dan $25.00 bagi sejuta token output.

- Pembangun: Anthropic
- Id: claude-opus-4-8
- Menurut laporan: 2026-06-12T00:00:00.000Z
- Dikeluarkan: 2026-05-28T00:00:00.000Z
- Pengesanan: 2026-09-21T18:47:57.048Z
  - [pautan bukti](https://models.dev/#claude-opus-4-8)
- Status: Dikeluarkan
- Kanonikal: https://alexgetman.com/ms/models/anthropic/claude-opus-4-8/

## Soalan

### Apakah tetingkap konteks Claude Opus 4.8?

Tetingkap konteks ialah 1,000,000 token. Keluaran maksimum ialah 128,000 token.

### Berapakah harga Claude Opus 4.8?

Harganya ialah $5.00 bagi sejuta token input dan $25.00 bagi sejuta token output. Input dalam cache berharga $0.50 bagi sejuta token.

### Bilakah Claude Opus 4.8 dikeluarkan?

Pada 2026-05-28.

## Garis masa

- 2026-07-01T16:16:29.000Z · [Alex Getman](https://alexgetman.com/23/sonnet-5-looks-absurdly-priced/)
> Sonnet 5 looks absurdly priced
> 
> • Opus 4.8 Max - $1.80
> • Sonnet 5 Max - $2.29
> • Fable 5 - $2.75
> 
> Sonnet 5 is about 27% more expensive than Opus 4.8.
> 
> Why does the lower-tier model cost more than the flagship?
- 2026-07-01T19:08:12.000Z · [Alex Getman](https://alexgetman.com/25/ai-already-automates-16-of-freelance-tasks/)
> AI completes 16% of the freelance tasks in Remote Labor Index
> 
> CAIS updated the Remote Labor Index — a benchmark where AI agents complete real client projects: design, video, CAD, architecture, analytics, and web apps.
> 
> The work is reviewed by humans and compared against a professional deliverable that a paying client would accept.
> 
> New results:
> 
> • Claude Fable 5 — 16.1%
> • Claude Opus 4.8 — 8.3%
> • GPT-5.5 — 6.3%
> 
> Fable 5 is now the top model in RLI and almost 2x higher than Opus 4.8.
> 
> When the benchmark launched, the best result was 2.5%.
- 2026-07-10T16:34:15.332Z · [Alex Getman](https://alexgetman.com/53/grok-4-5-a-new-competitor-to-gpt-and-claude/)
> Grok 4.5: a new competitor to GPT and Claude
> 
> Posting this a little late, but I prepared something interesting. Grok 4.5 is temporarily available to test for free in Grok Build and Cursor.
> 
> The model runs at around 80 tokens per second, supports a context window of up to 500,000 tokens, and costs $2 per million input tokens and $6 per million output tokens. According to Musk, Grok 4.5 was trained from scratch on SpaceXAI’s own foundation and is not built on top of Kimi like Composer.
> 
> In terms of quality, Grok 4.5 comes close to GPT-5.6 Terra, while its pricing is closer to Luna. Among Anthropic models, it sits between Haiku and Sonnet, making it especially attractive in terms of price-to-pe
- 2026-07-16T17:56:24.898Z · [Alex Getman](https://alexgetman.com/79/release-kimi-k3/)
> ⚡️ RELEASE: KIMI K3
> 
> Moonshot just dropped its new flagship: 2.8T parameters, a 1M-token context window, native vision, and a focus on large projects, coding, games, and 3D. Full weights are expected in the next few days.
> 
> API: $3 / $15 per million input/output tokens. Pricing-wise, that puts it around Sonnet 5 or GPT-5.6 Terra, while Moonshot’s own benchmarks have it taking shots at Opus 4.8 and Sol.
> 
> So basically: Opus-class performance at Sonnet pricing. You can already try it for free on the Kimi website or through Kimi Code CLI.
- 2026-07-25T17:22:06.482Z · [Alex Getman](https://alexgetman.com/111/anthropic-said-it-cut-80-of-the-system-prompt-for-opus-5/)
> Anthropic said it cut 80% of the system prompt for Opus 5
> 
> I checked the claim:
> Opus 4.8: 130K characters
> Opus 5: 135K characters
> 
> So the prompt actually got longer. Maybe they meant some specific section inside the system prompt. Or maybe not.
- 2026-08-03T08:01:30.947Z · [Alex Getman](https://alexgetman.com/142/alibaba-released-qwen-3-8-max/)
> ⚡️ Alibaba released Qwen 3.8 Max
> 
> The new 2.4T parameter flagship benchmarks in the same league as GPT-5.6 Sol, Claude Opus 4.8 and Fable 5.
> 
> API pricing is $2/M input and $6/M output, roughly half the cost of GPT-5.6 Terra and Claude Sonnet.
> 
> Now it’s up to users to see whether Qwen can beat DeepSeek-V4-Flash and GPT-5.6 Luna on price to performance.
- 2026-08-26T15:47:17.819Z · [Alex Getman](https://alexgetman.com/358/glm-5-3-flash-is-out/)
> 🚨 GLM 5.3 FLASH IS OUT
> 
> The first natively multimodal model in the GLM 5 series. 320B total parameters, with just 18B active.
> 
> It beats GLM 5.2 across most benchmarks and gets close to Claude Opus 4.8 on coding and agentic tasks, while costing around 10x less.
> 
> And yes, ox alpha on OpenRouter and OpenCode was GLM 5.3 Flash.
- 2026-10-07T19:43:10.391Z · [Claude Code · changelog](https://code.claude.com/docs/en/changelog)
- 2026-10-07T19:43:10.391Z · [Claude Code · changelog](https://code.claude.com/docs/en/changelog)
- 2026-10-07T19:43:10.391Z · [Claude Code · changelog](https://code.claude.com/docs/en/changelog)
- 2026-10-07T19:43:10.391Z · [Claude Code · changelog](https://code.claude.com/docs/en/changelog)
- 2026-10-30T21:21:12.926Z · [Anthropic · SDK releases](https://platform.claude.com/docs/en/release-notes/overview)
> After removal, requests to claude-opus-4-8 with speed: "fast" will return an error
- 2026-11-12T20:23:08.041Z · [Anthropic · official articles](https://www.anthropic.com/claude-fable-and-mythos-5-2)
- 2026-11-13T17:12:46.898Z · [Anthropic · official articles](https://www.anthropic.com/claude-opus-5-5)
- 2026-11-13T17:22:56.107Z · [Anthropic · official articles](https://www.anthropic.com/news/claude-opus-5)
> In Claude.ai, Claude Code, and Claude Cowork, any flagged requests will fall back to Opus 4.8 by default

## Pautan bukti

- [editorial](https://alexgetman.com/111/anthropic-said-it-cut-80-of-the-system-prompt-for-opus-5/)
- [Alex Getman](https://alexgetman.com/142/alibaba-released-qwen-3-8-max/)
- [editorial](https://alexgetman.com/23/sonnet-5-looks-absurdly-priced/)
- [editorial](https://alexgetman.com/25/ai-already-automates-16-of-freelance-tasks/)
- [Alex Getman](https://alexgetman.com/358/glm-5-3-flash-is-out/)
- [editorial](https://alexgetman.com/53/grok-4-5-a-new-competitor-to-gpt-and-claude/)
- [Alex Getman](https://alexgetman.com/79/release-kimi-k3/)
- [anthropic](https://api.anthropic.com/v1/models)
- [Artificial Analysis · benchmarks](https://artificialanalysis.ai/models)
- [Claude Code · changelog](https://code.claude.com/docs/en/changelog)
- [models-dev](https://models.dev/#claude-opus-4-8)
- [models.dev · catalogue](https://models.dev/#claude-opus-4.8)
- [OpenCode · discovery](https://opencode.ai/data/anthropic/claude-opus-4-8)
- [OpenRouter](https://openrouter.ai/anthropic/claude-opus-4.8)
- [anthropic-pricing](https://platform.claude.com/docs/en/about-claude/pricing)
- [review](https://platform.claude.com/docs/en/models/opus-4-8/overview)
- [Anthropic · SDK releases](https://platform.claude.com/docs/en/release-notes/overview)
- [Vercel AI Gateway](https://vercel.com/ai-gateway)
- [VoxelBench · leaderboard](https://voxelbench.ai/leaderboard)
- [anthropic-articles](https://www.anthropic.com/claude-fable-and-mythos-5-2)
- [anthropic-articles](https://www.anthropic.com/claude-opus-5-5)
- [Arena · leaderboards](https://www.anthropic.com/news/claude-opus-4-8)
- [anthropic-articles](https://www.anthropic.com/news/claude-opus-5)
- [Claude Code · model ids](https://www.npmjs.com/package/@anthropic-ai/claude-code/v/2.1.317)

## Isyarat

- [Claude Opus 4.8 dikesan di Arena](https://alexgetman.com/ms/signals/2026/09/13/gpt-5-5-high-6/)
- [Claude Opus 4.8 dikesan di Arena](https://alexgetman.com/ms/signals/2026/09/11/claude-opus-4-8-max/)