VerdictPal
Join

Models / Anthropic

Claude Opus 4.7

Claude Opus · Released 2026-04-16

Anthropic's April 2026 flagship. 13% coding lift over Opus 4.6, 3x more production tasks resolved, high-res vision up to 3.75 MP, new tokenizer. $5/$25 per 1M tokens, 1M context window.

Most capable generally available Claude model. Best for complex reasoning, long-running agentic tasks, and high-stakes code review.

#7 of 21 on Vals Index · current snapshot
66 Vals Index · 15 benchmark rows · 7 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite64
Context1M
Input / 1M$5.0
Output / 1M$25
Knowledge cutoff2026-02-01
Output speed61 t/s
TTFT (default API)16.72s
Statussolid
ReasoningCodingAgenticLong contextHuman preference
Reasoning
91
Coding
88
Agentic
69
Long context
70
Human preference
82

80average across 5 tested families

Benchmark placements

Where Claude Opus 4.7 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.

BenchmarkFamilyRankScoreSourceDate
Vending-Bench 2Agentic#1/ 5Andon Labs Vending-Bench 22026-06-01
SWE-bench VerifiedCoding#3/ 2388SWE-bench official leaderboard2026-05-30
ARC-AGIReasoning#5/ 1476ARC Prize leaderboard2026-05-15
LMArena (Chatbot Arena)Human preference#6/ 1382LMArena leaderboard2026-05-25
Vals IndexAgentic#7/ 2066Vals AI — Vals Index2026-06-04
Artificial Analysis Intelligence IndexReasoning#8/ 6754Artificial Analysis2026-07-21
SciCodeCoding#10/ 6555Artificial Analysis2026-07-21
GPQA DiamondReasoning#15/ 6991Artificial Analysis2026-07-21
AA time to first tokenPerformance#16/ 69Artificial Analysis2026-07-21
Humanity's Last ExamReasoning#18/ 6540Artificial Analysis2026-07-21
AA-LCRLong context#21/ 6470Artificial Analysis2026-07-21
Terminal-BenchAgentic#23/ 6969Terminal-Bench 2.0 leaderboard2026-05-20
τ³-Bench BankingAgentic#28/ 6429Artificial Analysis2026-07-21
IFBenchReasoning#42/ 5759Artificial Analysis2026-07-21
AA output speedPerformance#46/ 69Artificial Analysis2026-07-21
Family context

The three highest-scoring carded models in each capability family. Where Claude Opus 4.7 shows up, it's highlighted.

ReasoningClaude Opus 4.7 · 91
  1. 1
    Claude Mythos Preview
    95
  2. 2
    GPT-5.6 Sol
    94
  3. 3
    Kimi K3
    94
CodingClaude Opus 4.7 · 88
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticClaude Opus 4.7 · 69
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextClaude Opus 4.7 · 70
  1. 1
    Kimi K3
    75
  2. 2
    GPT-5.1
    75
  3. 3
    GPT-5.6 Sol
    74
Human preferenceClaude Opus 4.7 · 82
  1. 1
    GPT 5.5
    95
  2. 2
    Claude Opus 4.8
    92
  3. 3
    Gemini 3.5 Flash
    91
Newest receipts

Benchmark rows added to the public ledger for Claude Opus 4.7 in the last 120 days. Older rows live in the full table below.

15 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context70.33%Artificial Analysis2
AA output speedPerformance61 t/sArtificial Analysis2
AA time to first tokenPerformance16.72sArtificial Analysis2
Artificial Analysis Intelligence IndexReasoning53.5%Artificial Analysis2
GPQA DiamondReasoning91.4%Artificial Analysis2
Humanity's Last ExamReasoning39.6%Artificial Analysis2
IFBenchReasoning58.64%Artificial Analysis2
SciCodeCoding54.5%Artificial Analysis2
τ³-Bench BankingAgentic28.87%Artificial Analysis2
Vals IndexAgentic66.1%Vals AI — Vals Index49
Vending-Bench 2Agentic10936.76Andon Labs Vending-Bench 252
SWE-bench VerifiedCoding87.6%SWE-bench official leaderboard54
LMArena (Chatbot Arena)Human preference1412LMArena leaderboard59
Terminal-BenchAgentic69.4%Terminal-Bench 2.0 leaderboard64
ARC-AGIReasoning75.8%ARC Prize leaderboard69
Family coverage

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.

Reasoning91
Coding88
Agentic69
Long context70
Human preference82
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover or tab any dot for name, score, and input price.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMiniMaxxAIMetaZhipuXiaomiMeituanCohere

Full benchmark ledger

Every catalog benchmark for Claude Opus 4.7. Scores link to the original source; gaps mean no public row exists yet.

15 sourced rows

15 of 32 catalog benchmarks have a sourced row for Claude Opus 4.7.

Reasoning

Coding

Agentic

Long context

Performance

Human preference

Changelog
  1. Released Claude Opus 4.7 — 13% coding lift, 3x production tasks, 3.75 MP vision.
Sources