vs
0:10
overall capability wins (0 ties across 10 scored). o3 leads overall.
- Reasoning0:4
- Coding0:3
- Arena0:1
- Pricing3:0
- Speed1:0
- Specs0:2
Verdict
o3 leads coding (widest gap: +48.4 on Aider); o3 ranks higher on LMArena (+95 Elo); o3 offers ~1.6x larger context (200K vs 128K).
- o3 leads coding (widest gap: +48.4 on Aider)
- o3 ranks higher on LMArena (+95 Elo)
- o3 offers ~1.6x larger context (200K vs 128K)
- o3 allows ~1.5x more max output (100K vs 66K)
- o1 Mini is ~1.8x cheaper on a blended token basis than o3
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- o1 Mini
- $2.20
- o3
- $4
o1 Mini is about 1.8x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
OpenAI
OpenAI
Family
GPT-4 / o-series
GPT-4 / o-series
License
Proprietary
Proprietary
Open weights
No
No
Release
Sep 12, 2024
Apr 16, 2025
Knowledge cutoff
-
-
API / provider
OpenAI
OpenAI
Modalities
text → text
text, image → text
Specs
Context window
128K
200K
Max output
66K
100K
Parameters
—
—
Pricing
Input $/1M
$1.10
$2
Output $/1M
$4.40
$8
Blended $/1M (3∶1)
$1.93
$3.50
Speed
tok/s
90
55
TTFT (s)
1.2
1.2
Reasoning
MMLU-Pro
74.2
85.3
GPQA Diamond
60.3
82.7
Humanity's Last Exam
3.6
20.1
AIME 2025
—
88.9
MATH-500
94.4
99.2
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
38.6
AA-LCR v1.1
—
74.7
CritPt
—
1.1
MMMU-Pro
—
70.1
IFBench
—
71.4
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
58.4
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
57.6
80.8
Terminal-Bench 2.1
—
—
Aider Polyglot
32.9
81.3
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
32.3
41.0
CursorBench
—
—
SWE-Rebench
—
38.1
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,337
1,432
| Metric | o1 Mini | o3 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | OpenAI | — |
| Family | GPT-4 / o-series | GPT-4 / o-series | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Sep 12, 2024 | Apr 16, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | OpenAI | OpenAI | — |
| Modalities | text → text | text, image → text | — |
| Specs | |||
| Context window | 128K | 200K | B +72K |
| Max output | 66K | 100K | B +34K |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $1.10 | $2 | A +$0.90 |
| Output $/1M | $4.40 | $8 | A +$3.60 |
| Blended $/1M (3∶1) | $1.93 | $3.50 | A +$1.57 |
| Speed | |||
| tok/s | 90 | 55 | A +35 |
| TTFT (s) | 1.2 | 1.2 | tie |
| Reasoning | |||
| MMLU-Pro | 74.2 | 85.3 | B +11.1 pts |
| GPQA Diamond | 60.3 | 82.7 | B +22.4 pts |
| Humanity's Last Exam | 3.6 | 20.1 | B +16.5 pts |
| AIME 2025 | — | 88.9 | — |
| MATH-500 | 94.4 | 99.2 | B +4.8 pts |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 38.6 | — |
| AA-LCR v1.1 | — | 74.7 | — |
| CritPt | — | 1.1 | — |
| MMMU-Pro | — | 70.1 | — |
| IFBench | — | 71.4 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | 58.4 | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 57.6 | 80.8 | B +23.2 pts |
| Terminal-Bench 2.1 | — | — | — |
| Aider Polyglot | 32.9 | 81.3 | B +48.4 pts |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 32.3 | 41.0 | B +8.7 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | 38.1 | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,337 | 1,432 | B +95 Elo |
o1 Mini
- Context
- 128K / 66K out
- Parameters
- —
- Price
- $1.10 / $4.40
- Speed
- 90 tok/s · 1.2s TTFT
- Modalities
- text → text
- License
- Proprietary
Cost-efficient o1 reasoning model optimized for STEM and coding before the o3 / o4-mini line.
o3
- Context
- 200K / 100K out
- Parameters
- —
- Price
- $2 / $8
- Speed
- 55 tok/s · 1.2s TTFT
- Modalities
- text, image → text
- License
- Proprietary
OpenAI reasoning model from the o-series; strong on STEM and agentic coding with deliberate chain-of-thought.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.