Valumigo

Claude Opus 5

Anthropic · 2026-07-24

Basic Information

Developer
AnthropicAnthropic
Release Date
2026-07-24
Overall Capability Index (ECI)
162.9

Benchmark Scores (Epoch AI)

Each benchmark also shows the model's rank among the models evaluated.

AnthropicPhD-level science questions (GPQA Diamond) · #12 of 30
93.9%
AnthropicAdvanced mathematics (FrontierMath) · #11 of 30
85.6%
AnthropicReasoning through unfamiliar puzzles (ARC-AGI-2) · #4 of 30
90.4%
AnthropicFactual question accuracy (SimpleQA Verified) · #16 of 30
59.9%
AnthropicComputer interface tasks (OSWorld 2.0) · #1 of 9
31.4%

Bars show accuracy (%) and start at 0%.

User Vote Rankings (LMArena)

Rankings based on people's votes comparing responses from two models. Models do not appear on leaderboards where they have too few votes.

  • Overall#4 of 413 · 1504
  • English#6 of 413 · 1506
  • Web development#1 of 50 · 1815
  • Document understanding#1 of 44 · 1516
  • Creative writing#2 of 411 · 1516
  • Chinese#2 of 389 · 1571
  • Overall agent performance#2 of 50 · 0.138
  • Math#3 of 396 · 1521
  • Korean#3 of 278 · 1498
  • Japanese#4 of 274 · 1506
  • Agent task success#4 of 50 · 0.141
  • German#6 of 307 · 1506
  • Spanish#10 of 295 · 1484
  • Coding#12 of 408 · 1538
  • Image understanding (vision)#14 of 50 · 1289
  • French#17 of 290 · 1506

Where to Use It

Models from Anthropic are primarily available through Claude. Available models and usage limits vary by plan, so check the pricing and free access page.

View Claude pricing and free access →

Hands-on Test Reports

Responding to feedback that the first round was too easy, we tested bugs across multiple files, a feature defined by a specification, and a concurrency bug in a small online store codebase. We publish all 18 run records and task files.

Claude Code vs Codex, round 2: 3 practical tasks across multiple files, each run 3 times
Scores and rankings are reproduced directly from public sources and were not measured by this site. Model names vary slightly across sources and are matched automatically, so different versions may occasionally be mixed together.

Source: Epoch AI (CC BY 4.0) · Checked 2026-10-04 · LMArena (CC BY 4.0) · Checked 2026-10-04