Compare / Claude Opus 4.1 vs Claude Opus 4.5

Claude Opus 4.1 vs Claude Opus 4.5

Which is cheaper, which is more capable, and which does better at each kind of work, from 20 results on 4 sources that measured both models.

AnthropicClaude Opus 4.1 Released 5 Aug 2025 · $15.00 in · $75.00 out
AnthropicClaude Opus 4.5 Released 24 Nov 2025 · $5.00 in · $25.00 out
Claude Opus 4.1 wins 02 too close to callClaude Opus 4.5 wins 18

Of the 20 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.

Results as of 21 November 2026 (catalogue release 2026-11-21-af3a2faf6a5c).

Claude Opus 4.1 or Claude Opus 4.5? The short answer

Claude Opus 4.5 is the newer version and costs less, and it scores higher (Intelligence Index 29.1 against 22.8): upgrade.

  • Claude Opus 4.5 is cheaper: $10.00 against $30.00 per million tokens, 3.0 times less.
  • Claude Opus 4.5 scores higher on the Artificial Analysis Intelligence Index: 29.1 against 22.8.
  • Claude Opus 4.5 is newer: released 24 Nov 2025, against 5 Aug 2025 for Claude Opus 4.1.
  • On third-party benchmarks, on the results both models have: Claude Opus 4.5 does better for reasoning and knowledge, writing and sticking to the facts.

Head to head

22.8
Intelligence Index
29.1
$30.00
Price per million tokens, blendedlower is better
$10.00
200,000
Context window, tokens
200,000

Every fact side by side

Claude Opus 4.1Claude Opus 4.5
Price per million tokens$15.00 in · $75.00 out$5.00 in · $25.00 out
Blended (3 in : 1 out)$30.00$10.00
Cheapest host, blended–$10.00 (Amazon Bedrock)
Intelligence Index22.829.1
Spring Prompt overall––
Output speed, tokens a second––
Context window200,000 tokens200,000 tokens
Released5 Aug 202524 Nov 2025 (newer)
WeightsClosedClosed
Developer based inthe USthe US

Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.

Which is better for what

Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.

Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls not measured 36 of 159
Professional workReal tasks from banking, consulting and law, business documents and freelance projects not measured 28 of 51
Reasoning and knowledgeHard questions across science, maths and general knowledge 108 of 187 72 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 34 of 169 18 of 169
Sticking to the factsSummarising without inventing things, and factual answers 99 of 139 67 of 139

From our research

Every shared result

Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.

MeasureClaude Opus 4.1Claude Opus 4.5
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 69 · 33 of 387 measured 1,450thinking-16k reasoning 1,474high-32k reasoning
Overall · Text factualityArena rating, higher is better 88 · 41 of 163 measured 1,435 1,459
Business, management and finance · TextArena rating, higher is better 70 · 31 of 380 measured 1,447thinking-16k reasoning 1,472
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 92 · 60 of 249 measured 22.8thinking reasoning 29.1thinking reasoning
GPQA Diamond% of questions, higher is better 107 · 70 of 232 measured 80.9%thinking reasoning 86.6%thinking reasoning
Humanity's Last Exam% of questions, higher is better 145 · 80 of 250 measured 12.5%thinking reasoning 30.1%thinking reasoning
UGI Leaderboard reported by UGI Leaderboard · 3 results
Writing score · UGIscore out of 100, higher is better 10 · 8 of 230 measured 69.9no reasoning 70.4no reasoning
Requested-length error · UGI% off the requested word count, lower is better 8 · 8 of 230 measured 3.0%no reasoning 3.0%no reasoning
Style adherence · UGIscore from 0 to 1, higher is better 38 · 20 of 230 measured 0.38thinking reasoning 0.39thinking reasoning
Vectara Hallucination Leaderboard reported by Vectara Hallucination Leaderboard · 2 results
Hallucination rate · Vectara% of summaries, lower is better 75 · 69 of 104 measured 11.8% 10.9%
Answer rate · Vectara% of documents, higher is better 96 · 71 of 104 measured 92.4% 98.7%
Show all 20 results (9 more)
MeasureClaude Opus 4.1Claude Opus 4.5
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 5 results
Creative writing · TextArena rating, higher is better 40 · 13 of 385 measured 1,446thinking-16k reasoning 1,470high-32k reasoning
Expert prompts · TextArena rating, higher is better 59 · 33 of 338 measured 1,482thinking-16k reasoning 1,503high-32k reasoning
Instruction following · TextArena rating, higher is better 43 · 16 of 387 measured 1,458thinking-16k reasoning 1,483high-32k reasoning
Overall · SearchArena rating, higher is better 24 · 20 of 33 measured 1,167arena search 1,187arena search
Writing, literature and language · TextArena rating, higher is better 51 · 22 of 386 measured 1,446thinking-16k reasoning 1,467
Artificial Analysis reported by Artificial Analysis · 4 results
IFBench% of instructions, higher is better 90 · 84 of 195 measured 55.4%thinking reasoning 58.0%thinking reasoning
Long-context reasoning (AA-LCR)% of questions, higher is better 80 · 73 of 241 measured 76.0%thinking reasoning 77.3%thinking reasoning
Terminal-Bench Hard% of tasks, higher is better 55 · 15 of 190 measured 34.3%thinking reasoning 47.0%thinking reasoning
τ²-bench telecom% of tasks, higher is better 81 · 44 of 193 measured 71.4%thinking reasoning 89.5%thinking reasoning

Quick answers

Is Claude Opus 4.1 better than Claude Opus 4.5?

Claude Opus 4.5 is the newer version and costs less, and it scores higher (Intelligence Index 29.1 against 22.8): upgrade.

Which is cheaper, Claude Opus 4.1 or Claude Opus 4.5?

Claude Opus 4.5 is cheaper: $10.00 against $30.00 per million tokens, 3.0 times less.

Which is better for reasoning and knowledge, Claude Opus 4.1 or Claude Opus 4.5?

Claude Opus 4.5 ranks 72 of 187 models on sale for hard questions across science, maths and general knowledge, against 108 for Claude Opus 4.1.

Which is better for writing, Claude Opus 4.1 or Claude Opus 4.5?

Claude Opus 4.5 ranks 18 of 169 models on sale for what people prefer in blind comparisons, and judged writing quality, against 34 for Claude Opus 4.1.

Which is better for sticking to the facts, Claude Opus 4.1 or Claude Opus 4.5?

Claude Opus 4.5 ranks 67 of 139 models on sale for summarising without inventing things, and factual answers, against 99 for Claude Opus 4.1.

Run Claude Opus 4.1 and Claude Opus 4.5 on your own prompt.

Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.