o3 (2025-04-16)
Apr 2025
o3 (2025-04-16)Compare inon
Benchmarks of o3 (2025-04-16)31
BenchmarkByScoreRank
Knowledge15
Reasoning1
Math4
Legacy · saturated or no longer reported11
No benchmark matches
Rankingat high effort?
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperDeepSeek V4.1 Flash 1017maxOpen+69%-85%$0.26Pass back reasoning_content · Chat Completions+4
CheaperGemini 3.8 Flashhigh+62%-57%*$3.49Sampling parameters are ignored+1CheaperDeepSeek V4 Pro 0813maxOpen+48%-43%$0.67No image input+5
CheaperGrok 4.6xhigh+17%-14%$1.75
CheaperGLM 5.3maxOpen+14%-39%$2.01No image input+4ModelScoreCostPer taskWhat changes
DeepSeek V4.1 Flash 1017maxOpen+69%-85%$0.26Pass back reasoning_content · Chat Completions+4Gemini 3.8 Flashhigh+62%-57%*$3.49Sampling parameters are ignored+1DeepSeek V4 Pro 0813maxOpen+48%-43%$0.67No image input+5Grok 4.6xhigh+17%-14%$1.75GLM 5.3maxOpen+14%-39%$2.01No image input+4ModelScoreCostPer taskWhat changes
DeepSeek V4.1 Flash 1017maxOpen+69%-85%$0.26Pass back reasoning_content · Chat Completions+4DeepSeek V4 Pro 0813maxOpen+48%-43%$0.67No image input+5GPT-6.1 Solmax×2.8+14%$0.72Remove unsupported parameters+5GPT-6 Solmax+93%+14%$1.04GPT-5.6 Terramax+21%+29%$1.40Prompts with >272K input tokensModelScoreCostPer taskWhat changes
DeepSeek V4.1 Flash 1017maxOpen+69%-85%$0.26Pass back reasoning_content · Chat Completions+4DeepSeek V4 Pro 0813maxOpen+48%-43%$0.67No image input+5GLM 5.3maxOpen+14%-39%$2.01No image input+4
Kimi K3maxOpen-10%+71%$2.00Sampling parameters are fixed+3
Qwen3.6 35B A3BnoneOpen-24%-84%$0.25 / $1.49ModelScoreCostPer taskWhat changes
GPT-6 Astramax×2.9—$3.26Monitoring can return HTTP 403+2GPT-6.1 Solmax×2.8—$0.72Remove unsupported parameters+5GPT-5.6 Solmax×2—$1.99GPT-6 Solmax+93%—$1.04GPT-5.5high+79%—$1.40Prompts with >272K input tokensGPT-5.4 (2026-03-05)medium-3%—$2.50 / $15GPT-5.2 (2025-12-11)high-21%—$1.75 / $14GPT-5 (2025-08-07)high-21%—$1.25 / $10All OpenAI models →Benchmarks
all categoriesChat & Knowledge
SimpleQA Verified v149.4#29
MATH Level 597.8#3
FrontierMath18.7#22
FrontierMath Tier 42.1#23
FrontierMath Tiers 1–3 (v2)33.3#56
Chess Puzzles36#36
Mystery Game Puzzles29#26
GPQA Diamond · Epoch81.8#67
OTIS Mock AIME 2024 202583.9#67
FAQ
How much does o3 (2025-04-16) cost?›
$2 per 1M input tokens and $8 per 1M output tokens.
Is o3 (2025-04-16) free to use?›
There is no free API; the cheapest paid price is $2 per 1M input tokens.
How good is o3 (2025-04-16), and what is it best at?›
#91 of 205 in Chat & Knowledge, #20 of 24 in Coding, #4 of 39 in Agents & Tool Use, #91 of 187 in Writing & Language.
Is o3 (2025-04-16) better than Grok 4.7?›
On Mystery Game Puzzles, o3 (2025-04-16) scores 29 vs 29 — about the same.
How can I use o3 (2025-04-16)?›
Through the OpenAI API.
What is the context window of o3 (2025-04-16)?›
200K tokens.
When was o3 (2025-04-16) released? Is there a newer version?›
Released 2025-04-16. It has been superseded by GPT-5.6 Sol.
Can o3 (2025-04-16) read images?›
It accepts text and images as input.
What is o3 (2025-04-16)?›
o3 (2025-04-16) is a language model from OpenAI, released Apr 2025.
What are the best alternatives to o3 (2025-04-16)?›
DeepSeek V4.1 Flash 1017 (cheaper).