vs
11:2
overall capability wins (2 ties across 15 scored). o3 leads overall.
- Reasoning7:1
- Coding3:1
- Arena1:0
- Pricing0:3
- Speed0:2
Verdict
o3 leads coding (widest gap: +23.1 on Aider); o3 ranks higher on LMArena (+41 Elo); o4-mini is ~1.8x cheaper on a blended token basis than o3.
- o3 leads coding (widest gap: +23.1 on Aider)
- o3 ranks higher on LMArena (+41 Elo)
- o4-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.
- o3
- $4
- o4-mini
- $2.20
o4-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
Apr 16, 2025
Apr 16, 2025
Knowledge cutoff
-
-
API / provider
OpenAI
OpenAI
Modalities
text, image → text
text, image → text
Specs
Context window
200K
200K
Max output
100K
100K
Parameters
—
—
Pricing
Input $/1M
$2
$1.10
Output $/1M
$8
$4.40
Blended $/1M (3∶1)
$3.50
$1.93
Speed
tok/s
55
90
TTFT (s)
1.2
0.7
Reasoning
MMLU-Pro
85.3
—
GPQA Diamond
82.7
78.4
Humanity's Last Exam
20.1
16.5
AIME 2025
88.9
92.7
MATH-500
99.2
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
38.6
24.8
AA-LCR v1.1
74.7
61.0
CritPt
1.1
0.6
MMMU-Pro
70.1
69.2
IFBench
71.4
68.7
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
58.4
45.0
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
80.8
—
Terminal-Bench 2.1
—
—
Aider Polyglot
81.3
58.2
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
41.0
46.5
CursorBench
—
—
SWE-Rebench
38.1
27.5
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,432
1,391
| Metric | o3 | o4-mini | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | OpenAI | — |
| Family | GPT-4 / o-series | GPT-4 / o-series | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Apr 16, 2025 | Apr 16, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | OpenAI | OpenAI | — |
| Modalities | text, image → text | text, image → text | — |
| Specs | |||
| Context window | 200K | 200K | tie |
| Max output | 100K | 100K | tie |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $2 | $1.10 | B +$0.90 |
| Output $/1M | $8 | $4.40 | B +$3.60 |
| Blended $/1M (3∶1) | $3.50 | $1.93 | B +$1.57 |
| Speed | |||
| tok/s | 55 | 90 | B +35 |
| TTFT (s) | 1.2 | 0.7 | B +0.50 |
| Reasoning | |||
| MMLU-Pro | 85.3 | — | — |
| GPQA Diamond | 82.7 | 78.4 | A +4.3 pts |
| Humanity's Last Exam | 20.1 | 16.5 | A +3.6 pts |
| AIME 2025 | 88.9 | 92.7 | B +3.8 pts |
| MATH-500 | 99.2 | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 38.6 | 24.8 | A +13.8 pts |
| AA-LCR v1.1 | 74.7 | 61.0 | A +13.7 pts |
| CritPt | 1.1 | 0.6 | A +0.5 pts |
| MMMU-Pro | 70.1 | 69.2 | A +0.9 pts |
| IFBench | 71.4 | 68.7 | A +2.7 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | 58.4 | 45.0 | A +13.4 pts |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 80.8 | — | — |
| Terminal-Bench 2.1 | — | — | — |
| Aider Polyglot | 81.3 | 58.2 | A +23.1 pts |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 41.0 | 46.5 | B +5.5 pts |
| CursorBench | — | — | — |
| SWE-Rebench | 38.1 | 27.5 | A +10.6 pts |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,432 | 1,391 | A +41 Elo |
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.
o4-mini
- Context
- 200K / 100K out
- Parameters
- —
- Price
- $1.10 / $4.40
- Speed
- 90 tok/s · 0.7s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Cost-efficient o-series reasoning model for STEM and coding with lower latency than o3.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.