Models, measured side by side
GPT-5.2 Pro vs Hy3
Compare GPT-5.2 Pro, Hy3 by token pricing, context length, benchmark results, speed and tool support.
Monthly workload
Monthly cost breakdown
InputOutput
GPT-5.2 Pro
$1,260 / mo
Hy3
$5.28 / mo
Performance comparison
GPT-5.2 Pro
Not listed
Hy3
25.3
Overview
Pricing & capacity
| Pricing & capacity | GPT-5.2 Pro | Hy3 |
|---|---|---|
| Input / 1M tokens | $21 | $0.13 |
| Output / 1M tokens | $168 | $0.53 |
| Cached input / 1M | Not listed | $0.03 |
| Monthly cost | $1,260 | $5.28 |
| Context window | 400K | 262K |
| Maximum output | 128K | 128K |
Input / 1M tokens
- GPT-5.2 Pro
- $21
- Hy3
- $0.13
Output / 1M tokens
- GPT-5.2 Pro
- $168
- Hy3
- $0.53
Cached input / 1M
- GPT-5.2 Pro
- Not listed
- Hy3
- $0.03
Monthly cost
- GPT-5.2 Pro
- $1,260
- Hy3
- $5.28
Context window
- GPT-5.2 Pro
- 400K
- Hy3
- 262K
Maximum output
- GPT-5.2 Pro
- 128K
- Hy3
- 128K
Benchmarks & performance
| Benchmarks & performance | GPT-5.2 Pro | Hy3 |
|---|---|---|
| Arena Elo | 1,255 | Not listed |
| MMLU-Pro | 82.8% | Not listed |
| Intelligence Index | Not listed | 25.3 |
| GPQA | Not listed | 89.7% |
| Humanity’s Last Exam | Not listed | 33.5% |
| SciCode | Not listed | 48.6% |
| Output speed | Not listed | 88 tokens/s |
| Time to first token | Not listed | 3.08 s |
| DesignArena rating | Not listed | 1,180 |
| DesignArena win rate | Not listed | 41.1% |
Arena Elo
- GPT-5.2 Pro
- 1,255
- Hy3
- Not listed
MMLU-Pro
- GPT-5.2 Pro
- 82.8%
- Hy3
- Not listed
Intelligence Index
- GPT-5.2 Pro
- Not listed
- Hy3
- 25.3
GPQA
- GPT-5.2 Pro
- Not listed
- Hy3
- 89.7%
Humanity’s Last Exam
- GPT-5.2 Pro
- Not listed
- Hy3
- 33.5%
SciCode
- GPT-5.2 Pro
- Not listed
- Hy3
- 48.6%
Output speed
- GPT-5.2 Pro
- Not listed
- Hy3
- 88 tokens/s
Time to first token
- GPT-5.2 Pro
- Not listed
- Hy3
- 3.08 s
DesignArena rating
- GPT-5.2 Pro
- Not listed
- Hy3
- 1,180
DesignArena win rate
- GPT-5.2 Pro
- Not listed
- Hy3
- 41.1%
Tools & features
| Tools & features | GPT-5.2 Pro | Hy3 |
|---|---|---|
| Function calling | Supported | Supported |
| Structured outputs | Supported | Supported |
| JSON mode | Supported | Supported |
| Reasoning | Supported | Supported |
| Deterministic seed | Supported | Supported |
| Prompt caching | Not listed | Supported |
Function calling
- GPT-5.2 Pro
- Supported
- Hy3
- Supported
Structured outputs
- GPT-5.2 Pro
- Supported
- Hy3
- Supported
JSON mode
- GPT-5.2 Pro
- Supported
- Hy3
- Supported
Reasoning
- GPT-5.2 Pro
- Supported
- Hy3
- Supported
Deterministic seed
- GPT-5.2 Pro
- Supported
- Hy3
- Supported
Prompt caching
- GPT-5.2 Pro
- Not listed
- Hy3
- Supported
API & availability
| API & availability | GPT-5.2 Pro | Hy3 |
|---|---|---|
| API identifier | gpt-5.2-pro | hy3 |
| API providers | OpenAI | Not listed |
| Prices checked | Oct 8, 2026 | Oct 8, 2026 |
API identifier
- GPT-5.2 Pro
gpt-5.2-pro- Hy3
hy3
API providers
- GPT-5.2 Pro
- OpenAI
- Hy3
- Not listed
Prices checked
- GPT-5.2 Pro
- Oct 8, 2026
- Hy3
- Oct 8, 2026
About the models
GPT-5.2 Pro
GPT-5.2 with extra compute for the most demanding problems. No cached-input discount.
Full pricing & detailsHy3
Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...
Full pricing & detailsComparison FAQ
GPT-5.2 Pro: $21 input and $168 output per million tokens. Hy3: $0.13 input and $0.53 output per million tokens. The cheaper choice depends on your input-to-output ratio; a model can have a lower input rate but a higher output rate.