Skip to content
aiqbench

Anthropic · model passport

Anthropic: Claude Opus 4.1

claude-opus-4-1
  • Cloud only
  • Proprietary licence
  • 200,000 context
  • Released 2025-08-05(curated)
14/27axes covered
Overall · LMArena
1,450#71 of 396
Best percentile
88.4Instruction following
Cohort standing
55.9median 51.6

What the data says

Benchmarked on 14 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

  • 88.4 pct — Instruction following (1,458)
  • 87 pct — Coding (1,511)
  • 85.4 pct — Hard prompts (1,480)

Uneven: 58.9 percentile points between its best and worst axis — 88.4 pct on Instruction following versus 29.5 pct on HHEM Hallucination (Vectara).

  • Instruction following88.4 pct1,458Verify →
  • Coding87 pct1,511Verify →
  • Hard prompts85.4 pct1,480Verify →
  • Overall82.3 pct1,450Verify →
  • SimpleBench69 pct60.0%Verify →
  • GPQA Diamond58.4 pct77.3%Verify →
  • AIME (math)49.5 pct68.9%Verify →
  • Factuality Arena45.9 pct1,435Verify →
  • Terminal-Bench45.5 pct38.0%Verify →
  • Humanity's Last Exam43.5 pct11.5%Verify →
  • FrontierMath40.6 pct7.2%Verify →
  • SWE-bench Verified40.6 pct73.3%Verify →
  • WebDev Arena35.3 pct1,390Verify →
  • HHEM Hallucination (Vectara)29.5 pct88.2%Verify →
  • Vision ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • IFBench (Artificial Analysis)Not benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Use it hosted

API model IDanthropic/claude-opus-4.1
Context length200,000 tokens
Max output32,000 tokens
Acceptsimage, text, file
Input price$15 per 1M tokens
Cached input price$1.50 per 1M tokens
Output price$75 per 1M tokens

API parameters this endpoint supports

Include reasoningMax tokensReasoningStop sequencesTemperatureTool choiceTool callingTop-kTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"anthropic/claude-opus-4.1","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • claude-opus-4-1-20250805-thinking-16k stands at 55.9 (raw mean 57.2) — measured on 14 of 27 axes.

    standingtemplateVerify →

  • claude-opus-4-1-20250805-thinking-16k stands at 55.9 versus a cohort median of 51.6 across 145 models released within 180 days — above the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →