AA-OmniscienceVerified
48.8%
AA-Omniscience Accuracy
Models / Anthropic
Claude Opus 4.8
Claude Opus · Released 2026-05-28
Anthropic's May 2026 flagship.
Front-door model for hard reasoning, long documents, agentic coding, and tool-heavy research. The model we reach for first when a task is too hard for Sonnet but small enough to fit in one prompt.
Composite48
Context1M
Input / 1M$15
Output / 1M$75
Knowledge cutoff2026-02-01
Output speed58 t/s
Statusflagship
Family profile
Best published score in each covered benchmark family.
51/100 avg
- Knowledge
- 49
- Reasoning
- 80
- Math
- 47
- Coding
- 59
- Agentic
- 11
- Long context
- 58
6 tested benchmark families
Editor's note
Opus 4.8 is the model we reach for first when a task is hard enough to make a Sonnet hallucinate but small enough to fit in one prompt. The single-number rank (Vals, AA) is the highest we've measured.Benchmark placements
Per-benchmark positions
Where Claude Opus 4.8 places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.
| Benchmark | Family | Rank | Score | Source | Date |
|---|---|---|---|---|---|
| Terminal-Bench-Science | Agentic | #4/ 9 | 11 | Terminal-Bench-Science 0.1 announcement | 2026-08-27 |
| DeepSWE | Coding | #5/ 8 | 59 | DeepSWE leaderboard v1.1 | 2026-06-20 |
| FrontierMath | Math | #5/ 15 | 47 | Epoch AI FrontierMath Tiers 1-3 (v2) CSV | 2026-08-15 |
| AA-Omniscience | Knowledge | #9/ 12 | 49 | Artificial Analysis — AA-Omniscience | 2026-09-01 |
| CritPt | Reasoning | #10/ 12 | 21 | Artificial Analysis — CritPt | 2026-09-01 |
| Humanity's Last Exam | Reasoning | #11/ 186 | 49 | Artificial Analysis | 2026-09-28 |
| SciCode | Coding | #27/ 97 | 54 | Artificial Analysis | 2026-09-28 |
| AA output speed | Performance | #106/ 195 | Artificial Analysis | 2026-10-12 | |
| GPQA Diamond | Reasoning | #140/ 184 | 80 | GPQA Diamond paper leaderboard | 2026-05-15 |
| AA-LCR | Long context | #154/ 184 | 58 | Artificial Analysis — AA-LCR | 2026-06-17 |
Family context
Strongest rows in each family
The three highest-scoring models with pages in each capability family. Where Claude Opus 4.8 shows up, it's highlighted.
KnowledgeClaude Opus 4.8 · 49
- 167
- 265
- 361
ReasoningClaude Opus 4.8 · 80
- 196
- 295
- 395
MathClaude Opus 4.8 · 47
- 152
- 252
- 350
CodingClaude Opus 4.8 · 59
- 174
- 273
- 370
AgenticClaude Opus 4.8 · 11
- 191
- 289
- 389
Long contextClaude Opus 4.8 · 58
- 189
- 286
- 385
Newest receipts
7 recentLast 120 days
Benchmark rows added to the public ledger for Claude Opus 4.8 in the last 120 days. Older rows live in the full table below.
| Benchmark | Family | Score | Source | Days ago |
|---|---|---|---|---|
| AA output speed | Performance | 58 t/s | Artificial Analysis | 10 |
| Humanity's Last Exam | Reasoning | 48.7% | Artificial Analysis | 24 |
| SciCode | Coding | 54.4% | Artificial Analysis | 24 |
| CritPt | Reasoning | 20.9% | Artificial Analysis — CritPt | 51 |
| AA-Omniscience | Knowledge | 48.8% | Artificial Analysis — AA-Omniscience | 51 |
| Terminal-Bench-Science | Agentic | 10.5% | Terminal-Bench-Science 0.1 announcement | 56 |
| FrontierMath | Math | 47.24% | Epoch AI FrontierMath Tiers 1-3 (v2) CSV | 68 |
Family coverage
Best score per family
A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.
Price vs. performance
Where Claude Opus 4.8 sits on the scatter
Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.
AnthropicOpenAIGoogleAlibabaMetaZhipuxAIDeepSeekMiniMaxXiaomiMoonshotOtherMeituanMistralNVIDIACohere
Full benchmark ledger
Every catalog benchmark for Claude Opus 4.8. Scores link to the original source; gaps mean no public row exists yet.
10 of 13 catalog benchmarks have a sourced row for Claude Opus 4.8.
Knowledge
Reasoning
CritPtVerified
20.9%
CritPt composite (AA run)
GPQA DiamondVerified
79.8%
GPQA Diamond accuracy (CoT)
Humanity's Last ExamVerified
48.7%
HLE (AA run)
Math
FrontierMathVerified
47.24%
FrontierMath Tiers 1-3 (v2) best Epoch run
Coding
Agentic
Terminal-Bench-ScienceVerified
10.5%
TB-Science 0.1 resolution (Claude Code)
Long context
AA-LCRVerified
58.3%
AA-LCR pass@1
Performance
AA output speedVerified
58 t/s
Median output tokens/s (1k prompt, default provider)
Changelog
Version history
- Released Claude Opus 4.8 — top of Vals (70.17%) and AA (61.4).
- Claude Opus 4.7 retired for general availability.
Sources
Where the numbers come from
- Anthropic pricing· checked 2026-06-01
- Claude 4 family announcement· checked 2026-06-01
- Artificial Analysis· checked 2026-06-01
Questions this model page answers
Every answer below is assembled from the dated fields on this page. Nothing is written separately for search.
What is Claude Opus 4.8?
Claude Opus 4.8 is a Anthropic model: Anthropic's May 2026 flagship.
What is Claude Opus 4.8 for?
Front-door model for hard reasoning, long documents, agentic coding, and tool-heavy research. The model we reach for first when a task is too hard for Sonnet but small enough to fit in one prompt.
What does Claude Opus 4.8 cost?
Claude Opus 4.8 lists $15 input and $75 output per 1M tokens. Pricing last checked 2026-06-01.
How large is the Claude Opus 4.8 context window?
Claude Opus 4.8 publishes a 1M token context window.
What is the VerdictPal status of Claude Opus 4.8?
Claude Opus 4.8 carries Flagship status on the model atlas.
The Pack · Editorial newsletter
Know when the rules move.
One short email when a university changes its AI rule for theses, or a research tool changes its pricing or privacy terms. No tracking, no third-party analytics. Unsubscribe in one click.