MModelspectraIndependent AI Model Intelligence
Model profile · Updated 2026-09-08
Model Profile · Anthropic · US

Claude Opus 4.8

Best forDeep reasoning, Long documents, Coding
LicensingClosed API
Updated2026-09-08
01

Snapshot

Overall
81
rank #5 of 22
Coding
87
rank #4 of 22
Multimodal
89
rank #6 of 22
Input / Output
$5 / $25
USD per 1M tokens
Context
1M
usable recall 96
Speed
~40 tok/s
TTFT 1.0s

Aggregated from public sources and independently weighted; methodology on the Terms page. Scores within 3 points are treated as statistically tied.

02

Specs & pricing

Every tracked field. List API prices in USD per 1M tokens; verify current vendor pricing before purchase.

VendorAnthropic (US)
Released / current2026.05
LicensingClosed / managed API
Context window1M
Max output128K
Effective-context score96
Input / output price$5 / $25 per 1M tokens
Cache discountPrompt cache 90% off
Free tierNo free tier
Speed / TTFT~40 tok/s / 1.0s
Function calling88
Refusal rate~9%
English / Chinese94 / 78
Modalitiestext, image
Fine-tuningNo
Private deploymentNo
SOC2 / no-trainyes / yes
03

Strengths & watch-outs

Strengths

  • Previous flagship still capable
  • Well-proven stability

Watch-outs

  • Superseded by Opus 5
  • Same price — prefer the newer model
04

Where it ranks among 22 models

Overall 81/100 (#5), coding 87 (#4), multimodal 89 (#6). See the full boards and side-by-side compare on the leaderboard.
05

Pricing in practice

Illustrative monthly bill for 100M input + 30M output tokens: $1,250 at list; about $845 with 90% of inputs cache-hit. Your mix and cache-hit ratio change this.

Illustrative model using public list rates before any enterprise agreement; cache discount: Prompt cache 90% off.

06

Frequently asked questions

How much does Claude Opus 4.8 cost?
List API pricing is $5 input and $25 output per 1M tokens. Prompt cache 90% off. No free tier
What is Claude Opus 4.8's context window, and is it usable end to end?
It advertises a 1M window with 128K max output; its effective-context score is 96/100, which is the better predictor of whether long-document details are actually retained.
Is Claude Opus 4.8 good for coding?
Its coding aggregate is 87/100, rank #4 of 22. That puts it in the top tier for agentic, multi-file engineering.
Can I self-host Claude Opus 4.8?
No. It is a closed managed API; there is no weights download.
How strong is it in Chinese and on multimodal inputs?
Chinese score is 78 versus English 94; multimodal aggregate is 89/100 (rank #6), accepting text, image.