OpenAI model product
o3 mini
o3 mini is a smaller variant of O3 from OpenAI, expected to provide enhanced multimodal capabilities, improved reasoning, and more efficient resource utilization.
Updated Sep 8, 2026. Default version: o3-mini
LLMBoard Score35.1o3-mini
Coverage80%25 benchmark families
Context window200KTokens
Official input price$1.1OpenAI API
o3 mini Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
o3 mini Benchmark Results
Benchmark scores for o3-mini.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkLiveBench | Score84.60% | Rank01 | Participants38 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkMATH | Score97.90% | Rank01 | Participants71 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkMultilingual MMLU | Score80.70% | Rank01 | Participants5 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkCOLLIE | Score98.70% | Rank02 | Participants10 | Percentile88.89% | EvidenceC | Evaluated |
| BenchmarkMGSM | Score92.00% | Rank02 | Participants31 | Percentile96.67% | EvidenceC | Evaluated |
| BenchmarkAider-Polyglot Edit | Score60.40% | Rank03 | Participants10 | Percentile77.78% | EvidenceC | Evaluated |
| BenchmarkInternal API instruction following (hard) | Score50.00% | Rank03 | Participants7 | Percentile66.67% | EvidenceC | Evaluated |
| BenchmarkMulti-IF | Score79.50% | Rank03 | Participants23 | Percentile90.91% | EvidenceC | Evaluated |
| BenchmarkSWE-Lancer | Score18.00% | Rank04 | Participants4 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkIFEval | Score93.90% | Rank05 | Participants68 | Percentile94.03% | EvidenceC | Evaluated |
| BenchmarkComplexFuncBench | Score17.60% | Rank06 | Participants7 | Percentile16.67% | EvidenceC | Evaluated |
| BenchmarkSWE-Lancer (IC-Diamond subset) | Score7.40% | Rank06 | Participants6 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkGraphwalks parents <128k | Score58.30% | Rank07 | Participants11 | Percentile40.00% | EvidenceC | Evaluated |
| BenchmarkAider-Polyglot | Score66.70% | Rank09 | Participants22 | Percentile61.90% | EvidenceC | Evaluated |
| BenchmarkGraphwalks BFS <128k | Score51.00% | Rank09 | Participants11 | Percentile20.00% | EvidenceC | Evaluated |
| BenchmarkOpenAI-MRCR: 2 needle 128k | Score18.70% | Rank09 | Participants9 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkAIME 2024 | Score87.30% | Rank12 | Participants53 | Percentile78.85% | EvidenceC | Evaluated |
| BenchmarkFrontierMath | Score9.20% | Rank15 | Participants17 | Percentile12.50% | EvidenceC | Evaluated |
| BenchmarkTAU-bench Airline | Score32.40% | Rank21 | Participants23 | Percentile9.09% | EvidenceC | Evaluated |
| BenchmarkTAU-bench Retail | Score57.60% | Rank21 | Participants25 | Percentile16.67% | EvidenceC | Evaluated |
| BenchmarkMulti-Challenge | Score39.90% | Rank23 | Participants29 | Percentile21.43% | EvidenceC | Evaluated |
| BenchmarkMMLU | Score86.90% | Rank27 | Participants101 | Percentile74.00% | EvidenceC | Evaluated |
| BenchmarkSimpleQA | Score15.00% | Rank38 | Participants47 | Percentile19.57% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Verified | Score49.30% | Rank96 | Participants113 | Percentile15.18% | EvidenceC | Evaluated |
| BenchmarkGPQA | Score77.20% | Rank107 | Participants247 | Percentile56.91% | EvidenceC | Evaluated |
| BenchmarkLM Arena Text Style Control | Score1,363.35 rating | Rank135 | Participants210 | Percentile35.89% | EvidenceA | Evaluated |
| BenchmarkLM Arena Text | Score1,336.38 rating | Rank144 | Participants210 | Percentile31.58% | EvidenceA | Evaluated |
o3 mini Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenatext style control | Categorymath | Rank103 | Rating / score1,405.18 | Votes1,909 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry mathematical | Rank107 | Rating / score1,411.68 | Votes1,711 | ObservationsN/A | Result date |
| Arenatext style control | Categorykorean | Rank107 | Rating / score1,331.26 | Votes241 | ObservationsN/A | Result date |
| Arenatext | Categorykorean | Rank109 | Rating / score1,313.46 | Votes241 | ObservationsN/A | Result date |
| Arenatext | Categorymath | Rank109 | Rating / score1,395.60 | Votes1,909 | ObservationsN/A | Result date |
| Arenatext | Categoryindustry mathematical | Rank110 | Rating / score1,398.34 | Votes1,711 | ObservationsN/A | Result date |
| Arenatext style control | Categoryjapanese | Rank110 | Rating / score1,315.52 | Votes374 | ObservationsN/A | Result date |
| Arenatext | Categoryjapanese | Rank115 | Rating / score1,285.82 | Votes374 | ObservationsN/A | Result date |
| Arenatext style control | Categorychinese | Rank120 | Rating / score1,411.86 | Votes1,181 | ObservationsN/A | Result date |
| Arenatext style control | Categorycoding | Rank121 | Rating / score1,434.66 | Votes2,596 | ObservationsN/A | Result date |
| Arenatext style control | Categoryexpert | Rank123 | Rating / score1,398.08 | Votes847 | ObservationsN/A | Result date |
| Arenatext style control | Categoryhard prompts english | Rank124 | Rating / score1,413.36 | Votes2,944 | ObservationsN/A | Result date |
| Arenatext | Categoryexpert | Rank125 | Rating / score1,364.26 | Votes847 | ObservationsN/A | Result date |
| Arenatext style control | Categoryhard prompts | Rank126 | Rating / score1,401.18 | Votes4,386 | ObservationsN/A | Result date |
| Arenatext | Categoryhard prompts | Rank127 | Rating / score1,365.36 | Votes4,386 | ObservationsN/A | Result date |
| Arenatext style control | Categoryinstruction following | Rank127 | Rating / score1,366.18 | Votes6,681 | ObservationsN/A | Result date |
| Arenatext style control | Categoryfrench | Rank129 | Rating / score1,371.24 | Votes203 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry software and it services | Rank129 | Rating / score1,410.62 | Votes4,338 | ObservationsN/A | Result date |
| Arenatext | Categorychinese | Rank130 | Rating / score1,378.50 | Votes1,181 | ObservationsN/A | Result date |
| Arenatext | Categorycoding | Rank130 | Rating / score1,378.48 | Votes2,596 | ObservationsN/A | Result date |
| Arenatext | Categoryhard prompts english | Rank131 | Rating / score1,375.53 | Votes2,944 | ObservationsN/A | Result date |
| Arenatext style control | Categorylonger query | Rank131 | Rating / score1,372.70 | Votes2,160 | ObservationsN/A | Result date |
| Arenatext style control | Categorypolish | Rank131 | Rating / score1,327.02 | Votes2,635 | ObservationsN/A | Result date |
| Arenatext | Categoryinstruction following | Rank132 | Rating / score1,337.48 | Votes6,681 | ObservationsN/A | Result date |
| Arenatext | Categorypolish | Rank132 | Rating / score1,293.79 | Votes2,635 | ObservationsN/A | Result date |
| Arenatext | Categorygerman | Rank133 | Rating / score1,302.79 | Votes377 | ObservationsN/A | Result date |
| Arenatext style control | Categorygerman | Rank133 | Rating / score1,325.13 | Votes377 | ObservationsN/A | Result date |
| Arenatext | Categoryfrench | Rank134 | Rating / score1,333.80 | Votes203 | ObservationsN/A | Result date |
| Arenatext | Categoryspanish | Rank134 | Rating / score1,320.45 | Votes126 | ObservationsN/A | Result date |
| Arenatext style control | Categoryexclude ties | Rank134 | Rating / score1,329.31 | Votes12,679 | ObservationsN/A | Result date |
o3 mini Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $1.1 input, $4.4 output per 1M
- Official provider
- OpenAI
- Lowest third-party
- From $0.99 input, $4 output per 1M via Poe
- Tracked offerings
- 19
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderPoe | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$0.99 | Output / 1M$4 | Context200K | Updated |
| ProviderNanoGPT | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderImpossibl | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderJiekou.AI | Provider model IDo3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context131.1K | Updated |
| ProviderOpenRouter | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderOpenAI | Provider model IDo3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderLLM Gateway | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderNEAR AI Cloud | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderMerge Gateway | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderCloudflare AI Gateway | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderHelicone | Provider model IDo3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderAzure Cognitive Services | Provider model IDo3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderAzure | Provider model IDo3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderVercel AI Gateway | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderEden AI | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderDevPass (LLM Gateway) | Provider model IDo3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderKilo Gateway | Provider model IDopenai/o3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
| ProviderAbacus | Provider model IDo3-mini | Regionglobal | Input / 1M$1.1 | Output / 1M$4.4 | Context200K | Updated |
o3 mini Runtime Performance
Provider-specific output speed and catalog latency for o3-mini. Runtime does not affect the capability score.
Sort by
Provider | Output Speed | Catalog Latency | Max Input | Max Output | Updated |
|---|
| ProviderAzure | Output Speed115.00 tok/s | Catalog Latency5.20 s | Max Input200K | Max Output100K | Updated |
| ProviderOpenAI | Output Speed115.00 tok/s | Catalog Latency5.20 s | Max Input200K | Max Output100K | Updated |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
o3 mini Specifications
Technical details for the model's default version.
- Version
- o3-mini
- Released
- Jan 30, 2025
- Knowledge cutoff
- Sep 30, 2023
- Parameters
- N/A
- Context window
- 200K
- Max output
- 100K
- Inputs
- text
- Outputs
- text
- Open weights
- No
- License
- Proprietary
o3 mini Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| Versiono3-mini | Released | LLMBoard35.06 | ParametersN/A | Context200K | Max output100K | Open weightsNo | LicenseProprietary |
Models similar to o3 mini
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
OP
GPT-4.5
OpenAI
OP
o1
OpenAI
OP
o1 pro
OpenAI
OP
GPT-4.1
OpenAI
OP
GPT-5-Nano
OpenAI
OP
GPT-5.3
OpenAI
What is o3 mini?
Key information about o3 mini and its available data.
o3 mini is a large language model from OpenAI and a smaller variant of O3. It is expected to improve multimodal capabilities, reasoning, and resource efficiency while maintaining performance on core tasks.
Data as of 2026-09-08.
FAQ
Common questions about o3 mini.
When was o3 mini released?
o3 mini's default version was released on Jan 30, 2025.
How much does o3 mini cost?
o3 mini's official API price is $1.1 per million input tokens and $4.4 per million output tokens via OpenAI. The lowest tracked third-party offer starts at $0.99 input and $4 output via Poe.
Who created o3 mini?
o3 mini was created by OpenAI.
What is the context window for o3 mini?
The default version has a 200K token context window.
Is o3 mini open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer o3 mini?
19 provider offerings are linked to the default version.