Claude Opus 4.8

AnthropicKey claude-opus-4-8Also reported as router_client:claude/claude-opus-4-8, claude/claude-opus-4-8, kiro/claude-opus-4-8, cc/claude-opus-4-8, anthropic/claude-opus-4-8, kr/claude-opus-4-8, vibeproxy/claude-opus-4-8, mvx/claude-opus-4-8, stealth/claude-opus-4-8, teamclaude/claude-opus-4-8, xr/claude-opus-4-8, router_client:unknown/claude-opus-4-8, bebal/claude-opus-4-8, eco/claude-opus-4-8, t/claude-opus-4-8, kiroaas/claude-opus-4-8, customoai/kami/claude-opus-4-8, sekai/claude-opus-4-8, vclaude/claude-opus-4-8, nghi/claude-opus-4-8

Rank
#12
by active users
Active users
222
Down 85 from 307 active users in the previous 30 days. vs previous 30 days
Adoption
23%
of 961 accounts
Tokens
28.7B
cache included
Est. cost
$24.9K
at public list prices
First seen
Jan 3, 2026
in any submission

Oct 14 – Nov 12, 2026 (UTC)961 accounts100% of tokens from an identified modelData through Nov 12, 2026, 06:01 UTCHow we count

Daily active users

Distinct accounts that used this model on each UTC day.

3 active users on Nov 12, 2026

Daily active users for Claude Opus 4.8, Oct 14, 2026 to Nov 12, 2026. Peak 44. Use the left and right arrow keys to read each day.50250
Show daily values
Daily active users and tokens for Claude Opus 4.8
Date (UTC)Active usersTokens
Oct 14, 202641996.3M
Oct 15, 2026331.3B
Oct 16, 2026441.1B
Oct 17, 2026261.3B
Oct 18, 2026282.1B
Oct 19, 2026402.1B
Oct 20, 2026401.3B
Oct 21, 2026421.2B
Oct 22, 2026381.6B
Oct 23, 2026421.7B
Oct 24, 2026281.9B
Oct 25, 202623989M
Oct 26, 2026322.5B
Oct 27, 2026331.1B
Oct 28, 2026331.2B
Oct 29, 202628712.2M
Oct 30, 202628656.3M
Oct 31, 202617265.3M
Nov 1, 202619569.2M
Nov 2, 202628366.7M
Nov 3, 202621266.2M
Nov 4, 202622282.3M
Nov 5, 202627723.3M
Nov 6, 202626670.7M
Nov 7, 202613678.2M
Nov 8, 202614277.4M
Nov 9, 202612525.2M
Nov 10, 202611265.6M
Nov 11, 20261087M
Nov 12, 2026314.6M

Token mix

How the 28.7B tokens split. Cache reads are usually most of an agent's tokens and are priced far below input.

  • Input0.2%
  • Output0.5%
  • Cache read96%
  • Cache write3.4%
  • Reasoning<0.1%

Agents that carry Claude Opus 4.8

Each agent's share of this model's tokens, and how many accounts used the model through it.

  1. Claude Code89% of this model's tokens213active users
  2. Pi5.0% of this model's tokens5active users
  3. Oh My Pi4.1% of this model's tokens2active users
  4. Senpi (OmO Native)0.7% of this model's tokens6active users
  5. Prime Agent0.6% of this model's tokens1active users
  6. ZCode0.5% of this model's tokens1active users
  7. DeepSeek Harness0.2% of this model's tokens1active users
  8. OpenCode0.2% of this model's tokens5active users
  9. Hermes Agent<0.1% of this model's tokens3active users
  10. Trae<0.1% of this model's tokens1active users
  11. Codex CLI<0.1% of this model's tokens1active users
  12. Droid<0.1% of this model's tokens1active users
  13. Aside<0.1% of this model's tokens1active users

Benchmarks

The best result for Claude Opus 4.8 in each benchmark Tokscale tracks. Each benchmark has its own unit, run conditions and source, so compare a score only with other models on the same benchmark. Rank is among the models Tokscale matched in that benchmark. Agent rows also show Tokscale usage through that agent, side by side with the score. Usage does not explain the score.

All benchmarks
  • Epoch Capabilities IndexOverall · Epoch AI
    159.9
    Rank #21 of 158 matched modelsNo conditions reportedModel released May 28, 2026Source: epoch.ai (opens in a new tab)
  • Terminal-Bench 4.1Agentic coding · Terminal-Bench
    23.6%
    Rank #14 of 23 matched modelsMax · Claude CodePublished Nov 3, 2026Source: tbench.ai (opens in a new tab)

    Tokscale usage: Claude Code carried 89% of Claude Opus 4.8 tokens in the last 30 days (213 active users).

  • DeepSWEAgentic coding · Datacurve, collected by Epoch AI
    59.0%
    Rank #14 of 26 matched modelsMax · mini-swe-agentModel released May 28, 2026Source: deepswe.datacurve.ai (opens in a new tab)
    Show 4 more results with other run conditionsHide the other results
    • Listed as claude-opus-4-8 (xhigh)
      54.4%
      XHigh · mini-swe-agentModel released May 28, 2026Source: deepswe.datacurve.ai (opens in a new tab)
    • Listed as claude-opus-4-8 (high)
      51.8%
      High · mini-swe-agentModel released May 28, 2026Source: deepswe.datacurve.ai (opens in a new tab)
    • Listed as claude-opus-4-8 (medium)
      48.7%
      Medium · mini-swe-agentModel released May 28, 2026Source: deepswe.datacurve.ai (opens in a new tab)
    • Listed as claude-opus-4-8 (low)
      40.8%
      Low · mini-swe-agentModel released May 28, 2026Source: deepswe.datacurve.ai (opens in a new tab)
  • WebDev ArenaWeb development · LMArena
    1,556
    Rank #37 of 110 matched modelsHighPublished Oct 19, 2026Source: huggingface.co (opens in a new tab)
    Show 1 more result with other run conditionsHide the other results
    • Listed as Claude Opus 4.8
      1,536
      No conditions reportedPublished Oct 19, 2026Source: huggingface.co (opens in a new tab)
  • GPQA DiamondReasoning · Epoch AI
    91.0%
    Rank #35 of 160 matched modelsMaxRun date Jun 7, 2026Source: epoch.ai (opens in a new tab)

    Near its ceiling: the top score is 96.5%, and Epoch AI estimates the reachable maximum at 97%. Differences between top models are small.

    Show 2 more results with other run conditionsHide the other results
    • Listed as claude-opus-4-8_low
      88.4%
      LowRun date Aug 6, 2026Source: epoch.ai (opens in a new tab)
    • Listed as claude-opus-4-8_none
      85.4%
      No reasoningRun date Aug 6, 2026Source: epoch.ai (opens in a new tab)
  • FrontierMath Tiers 1–3 (v2)Reasoning · Epoch AI
    80.0%
    Rank #20 of 90 matched modelsMaxRun date Jun 10, 2026Source: epoch.ai (opens in a new tab)

    Near its ceiling: the top score is 95.1%, and Epoch AI estimates the reachable maximum at 95%. Differences between top models are small.

  • SimpleQA VerifiedKnowledge · Epoch AI
    53.0%
    Rank #26 of 86 matched modelsMaxRun date Aug 27, 2026Source: epoch.ai (opens in a new tab)

Not evaluated: Humanity's Last Exam, CursorBench, FrontierSWE, Agent Arena. A missing result is not a score of zero.

Pricing

List prices from models.dev in US dollars per 1M tokens, retrieved Nov 12, 2026. Anthropic is the lab's own API and the reference price for the estimates below.

Price at real agent usageEstimate at list prices

What 1M tokens cost at the reference list price when they split into input, output and cache the way Tokscale users' agents used them. It is not what anyone paid.

At this model's mix
$0.83
Est. per 1M tokens
96% cache reads
At the community mix
$0.76
Est. per 1M tokens
96% cache reads
3:1 input/output blend
$10
Per 1M tokens at list prices
No cache

Cache reads were 96% of Claude Opus 4.8 tokens in the last 30 days and 96% of all tokens. The 3:1 blend, (3 × input + output) ÷ 4, is the usual comparison and ignores cache.

Assumptions

  • Anthropic list prices per 1M tokens: input $5, output $25, cache read $0.50, cache write $6.25.
  • This model's mix is its own input, output, cache and reasoning tokens in the last 30 days.
  • The community mix is the tokens of every identified model in the last 30 days.
  • This is an estimate at list prices, not what anyone paid.
  • Reasoning tokens are priced as output.
  • Long-context tiers, batch pricing and discounts are not applied.
  • Token mixes come from public Tokscale submissions. Hidden accounts are excluded.

List prices by provider

  • AnthropicReferenceclaude-opus-4-8
    $5 input per 1M tokens$25 output per 1M tokens$0.50 cache read per 1M tokens$6.25 cache write per 1M tokens1M tokens of context
  • Amazon Bedrockanthropic.claude-opus-4-8
    $5 input per 1M tokens$25 output per 1M tokens$0.50 cache read per 1M tokens$6.25 cache write per 1M tokens1M tokens of context
  • Vertexclaude-opus-4-8@default
    $5 input per 1M tokens$25 output per 1M tokens$0.50 cache read per 1M tokens$6.25 cache write per 1M tokens1M tokens of context
  • Vertex (Anthropic)claude-opus-4-8@default
    $5 input per 1M tokens$25 output per 1M tokens$0.50 cache read per 1M tokens$6.25 cache write per 1M tokens1M tokens of context
Show 3 more providersShow fewer providers
  • Azureclaude-opus-4-8
    $5 input per 1M tokens$25 output per 1M tokens$0.50 cache read per 1M tokens$6.25 cache write per 1M tokens1M tokens of context
  • OpenRouteranthropic/claude-opus-4.8
    $5 input per 1M tokens$25 output per 1M tokens$0.50 cache read per 1M tokens$6.25 cache write per 1M tokens1M tokens of context
  • Vercel AI Gatewayanthropic/claude-opus-4.8
    $5 input per 1M tokens$25 output per 1M tokens$0.50 cache read per 1M tokens$6.25 cache write per 1M tokens1M tokens of context

Specs

From models.dev (MIT), retrieved Nov 12, 2026, as the lab or the first listed provider reports them. A dash means models.dev does not list the value.

Context window
1M tokens
Max output
128K tokens
Input
Text, Image, PDF
Output
Text
Knowledge cutoff
Jan 2026
Release date
May 28, 2026
Open weights
No
Reasoning
Yes
Tool calling
Yes

Top public profiles

Public profiles that used Claude Opus 4.8 most in this period, by tokens. Hidden accounts are never listed.

  1. 1Koala@koalaauto3Btokens$2.2Kest. cost11active days$2.2K est. cost · 11 active days
  2. 2Pandy@PandyOnGit1.9Btokens$1.3Kest. cost10active days$1.3K est. cost · 10 active days
  3. 3Alexander Alemayhu@aalemayhu1.7Btokens$1.2Kest. cost7active days$1.2K est. cost · 7 active days
  4. 4szymonsec@szymonsec1.5Btokens$1.6Kest. cost15active days$1.6K est. cost · 15 active days
  5. 5Jeff Emanuel@Dicklesworthstone1.3Btokens$946est. cost10active days$946 est. cost · 10 active days
  6. 6Shad0w23333@Shad0w233331.2Btokens$902est. cost17active days$902 est. cost · 17 active days
  7. 7Soju06@Soju06965.6Mtokens$1.1Kest. cost14active days$1.1K est. cost · 14 active days
  8. 8LJM@lee98www939.8Mtokens$591est. cost8active days$591 est. cost · 8 active days
  9. 9yunwang6131@yunwang6131759.6Mtokens$729est. cost11active days$729 est. cost · 11 active days
  10. 10G Man@genx7up695.7Mtokens$483est. cost14active days$483 est. cost · 14 active days

More from Anthropic

Other Anthropic models in the same period, by active users.

  1. Claude Opus 5.574% adoption713active users
  2. Claude Sonnet 5.558% adoption557active users
  3. Claude Haiku 5.550% adoption479active users
  4. Claude Fable 5.138% adoption362active users
  5. Claude Fable 5.220% adoption188active users
  6. Claude Opus 517% adoption164active users
All models

Sources and attribution

Each score keeps its own source, unit and run conditions. Tokscale does not combine benchmarks into one score, and a missing score means the model was not evaluated, not that it scored zero.

  • Epoch AIEpoch Capabilities Index, Epoch's own runs and the external results it collects.CC BY 4.0 · Retrieved Nov 12, 2026 · Data through Nov 9, 2026Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from https://epoch.ai/benchmarks.
  • LMArenaWebDev Arena and Agent Arena leaderboards.CC BY 4.0 · Retrieved Oct 21, 2026 · Data through Oct 19, 2026LMArena leaderboard dataset (lmarena-ai/leaderboard-dataset), Hugging Face.
  • Terminal-BenchTerminal-Bench 4.1 official leaderboard rows (results only, no task content).Terms not stated · Retrieved Nov 12, 2026 · Data through Nov 11, 2026Terminal-Bench 4.1 leaderboard, tbench.ai. Every score links back to the leaderboard.
  • models.devSpecs and list prices per provider.MIT · Retrieved Nov 12, 2026models.dev, an open catalog of AI models (github.com/anomalyco/models.dev).
  • TokscaleActive users, tokens and token mix from public Tokscale submissions. Hidden accounts are excluded.How we count