vs
0:2
overall capability wins (0 ties across 2 scored). o4-mini leads overall.
- Pricing0:3
- Speed0:2
- Specs0:2
Verdict
o4-mini offers ~6.1x larger context (200K vs 33K); o4-mini allows ~12.2x more max output (100K vs 8K); o4-mini is ~39.0x cheaper on a blended token basis than GPT-4 32K.
- o4-mini offers ~6.1x larger context (200K vs 33K)
- o4-mini allows ~12.2x more max output (100K vs 8K)
- o4-mini is ~39.0x cheaper on a blended token basis than GPT-4 32K
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- GPT-4 32K
- $90
- o4-mini
- $2.20
o4-mini is about 40.9x 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
Jun 13, 2023
Apr 16, 2025
Knowledge cutoff
-
-
API / provider
OpenAI
OpenAI
Modalities
text → text
text, image → text
Specs
Context window
33K
200K
Max output
8K
100K
Parameters
—
—
Pricing
Input $/1M
$60
$1.10
Output $/1M
$120
$4.40
Blended $/1M (3∶1)
$75
$1.93
Speed
tok/s
30
90
TTFT (s)
1
0.7
Reasoning
MMLU-Pro
—
—
GPQA Diamond
—
78.4
Humanity's Last Exam
—
16.5
AIME 2025
—
92.7
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
24.8
AA-LCR v1.1
—
61.0
CritPt
—
0.6
MMMU-Pro
—
69.2
IFBench
—
68.7
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
45.0
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
—
—
Aider Polyglot
—
58.2
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
46.5
CursorBench
—
—
SWE-Rebench
—
27.5
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
1,391
| Metric | GPT-4 32K | o4-mini | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | OpenAI | — |
| Family | GPT-4 / o-series | GPT-4 / o-series | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Jun 13, 2023 | Apr 16, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | OpenAI | OpenAI | — |
| Modalities | text → text | text, image → text | — |
| Specs | |||
| Context window | 33K | 200K | B +167K |
| Max output | 8K | 100K | B +92K |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $60 | $1.10 | B +$58.90 |
| Output $/1M | $120 | $4.40 | B +$115.60 |
| Blended $/1M (3∶1) | $75 | $1.93 | B +$73.08 |
| Speed | |||
| tok/s | 30 | 90 | B +60 |
| TTFT (s) | 1 | 0.7 | B +0.30 |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | — | 78.4 | — |
| Humanity's Last Exam | — | 16.5 | — |
| AIME 2025 | — | 92.7 | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 24.8 | — |
| AA-LCR v1.1 | — | 61.0 | — |
| CritPt | — | 0.6 | — |
| MMMU-Pro | — | 69.2 | — |
| IFBench | — | 68.7 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | 45.0 | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | — | — |
| Aider Polyglot | — | 58.2 | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | 46.5 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | 27.5 | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | 1,391 | — |
GPT-4 32K
- Context
- 33K / 8K out
- Parameters
- —
- Price
- $60 / $120
- Speed
- 30 tok/s · 1s TTFT
- Modalities
- text → text
- License
- Proprietary
Long-context GPT-4 SKU (gpt-4-32k-0613) at premium rates — predecessor to 128K Turbo for document workloads.
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.
These two models have no overlapping published benchmarks in our dataset. Compare specs and pricing instead.