Compare / o1 vs o3

o1 vs o3

Which is cheaper, which is more capable, and which does better at each kind of work, from 17 results on 4 sources that measured both models.

OpenAIo1 Released 17 Dec 2024 · Not on sale
OpenAIo3 Released 16 Apr 2025 · $2.00 in · $8.00 out
o1 wins 13 too close to callo3 wins 13

Of the 17 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.

Results as of 16 November 2026 (catalogue release 2026-11-16-49bdfa045dee).

o1 or o3? The short answer

o3 is the newer version at about the same price, and it scores higher (Intelligence Index 20.2 against 15.2): upgrade.

  • o3 scores higher on the Artificial Analysis Intelligence Index: 20.2 against 15.2.
  • o3 is newer: released 16 Apr 2025, against 17 Dec 2024 for o1.
  • On third-party benchmarks, on the results both models have: o3 does better for reasoning and knowledge, writing and sticking to the facts.

Head to head

15.2
Intelligence Index
20.2

Every fact side by side

o1o3
Price per million tokensNot on sale$2.00 in · $8.00 out
Blended (3 in : 1 out)–$3.50
Cheapest host, blended––
Intelligence Index15.220.2
Spring Prompt overall––
Output speed, tokens a second–155
Context window–200,000 tokens
Released17 Dec 202416 Apr 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 16 of 159
Reasoning and knowledgeHard questions across science, maths and general knowledge 126 of 188 105 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 72 of 170 52 of 169
Sticking to the factsSummarising without inventing things, and factual answers 77 of 140 88 of 139

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.

Measureo1o3
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 135 · 95 of 387 measured 1,402 1,432
Business, management and finance · TextArena rating, higher is better 166 · 99 of 380 measured 1,370 1,427
Creative writing · TextArena rating, higher is better 118 · 116 of 385 measured 1,381 1,383
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 133 · 113 of 249 measured 15.2 20.2
GPQA Diamond% of questions, higher is better 141 · 99 of 232 measured 74.7% 82.7%
Humanity's Last Exam% of questions, higher is better 178 · 117 of 250 measured 7.0% 20.1%
SimpleQA Verified (Epoch AI) reported by SimpleQA Verified (Epoch AI) · 1 result
Correct answers% of questions, higher is better 41 · 25 of 76 measured 41.1%high reasoning 49.4%high reasoning
UGI Leaderboard reported by UGI Leaderboard · 3 results
Writing score · UGIscore out of 100, higher is better 41 · 24 of 230 measured 61.0low reasoning 66.8medium reasoning
Requested-length error · UGI% off the requested word count, lower is better 140 · 91 of 230 measured 24.0%high reasoning 18.0%high reasoning
Style adherence · UGIscore from 0 to 1, higher is better 66 · 198 of 230 measured 0.36high reasoning 0.31high reasoning
Show all 17 results (7 more)
Measureo1o3
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Expert prompts · TextArena rating, higher is better 152 · 101 of 338 measured 1,402 1,449
Instruction following · TextArena rating, higher is better 113 · 120 of 387 measured 1,407 1,403
Writing, literature and language · TextArena rating, higher is better 110 · 114 of 386 measured 1,398 1,396
Artificial Analysis reported by Artificial Analysis · 4 results
IFBench% of instructions, higher is better 53 · 46 of 195 measured 70.3% 71.4%
Long-context reasoning (AA-LCR)% of questions, higher is better 134 · 86 of 241 measured 65.0% 74.7%
Terminal-Bench Hard% of tasks, higher is better 120 · 44 of 190 measured 12.9% 37.1%
τ²-bench telecom% of tasks, higher is better 94 · 69 of 193 measured 62.6% 80.7%

Quick answers

Is o1 better than o3?

o3 is the newer version at about the same price, and it scores higher (Intelligence Index 20.2 against 15.2): upgrade.

Which is better for reasoning and knowledge, o1 or o3?

o3 ranks 105 of 188 models on sale for hard questions across science, maths and general knowledge, against 126 for o1.

Which is better for writing, o1 or o3?

o3 ranks 52 of 170 models on sale for what people prefer in blind comparisons, and judged writing quality, against 72 for o1.

Which is better for sticking to the facts, o1 or o3?

o1 ranks 77 of 140 models on sale for summarising without inventing things, and factual answers, against 88 for o3.

Run o1 and o3 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.