vs
0:2
overall capability wins (0 ties across 2 scored). Hy4 Preview leads overall.
- Reasoning0:1
- Pricing3:0
- Specs0:1
Verdict
Hy4 Preview edges reasoning & knowledge; Hy4 Preview offers ~61.0x larger context (1M vs 16K); Hy4 Preview ships open weights (Apache 2.0).
- Hy4 Preview edges reasoning & knowledge
- Hy4 Preview offers ~61.0x larger context (1M vs 16K)
- Hy4 Preview ships open weights (Apache 2.0)
- GPT-3.5 Turbo is ~1.7x cheaper on a blended token basis than Hy4 Preview
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- GPT-3.5 Turbo
- $0.88
- Hy4 Preview
- $1.46
GPT-3.5 Turbo is about 1.7x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
OpenAI
Tencent
Family
Other
Other
License
Proprietary
Apache 2.0
Open weights
No
Yes
Release
Mar 1, 2023
Aug 28, 2026
Knowledge cutoff
-
-
API / provider
OpenAI
Tencent Cloud TokenHub
Modalities
text → text
text → text
Specs
Context window
16K
1M
Max output
4K
-
Parameters
—
770B (49B act.)
Pricing
Input $/1M
$0.50
$0.83
Output $/1M
$1.50
$2.50
Blended $/1M (3∶1)
$0.75
$1.25
Speed
tok/s
120
-
TTFT (s)
0.25
-
Reasoning
MMLU-Pro
46.2
—
GPQA Diamond
29.7
92.3
Humanity's Last Exam
—
—
AIME 2025
—
—
MATH-500
44.1
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
—
AA-LCR v1.1
—
—
CritPt
—
—
MMMU-Pro
—
—
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
65.7
SWE-bench Multilingual
—
82.9
LiveCodeBench
—
—
Terminal-Bench 2.1
—
85.4
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
—
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
64.3
WebDev Arena
—
1,624
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,225
—
| Metric | GPT-3.5 Turbo | Hy4 Preview | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | Tencent | — |
| Family | Other | Other | — |
| License | Proprietary | Apache 2.0 | — |
| Open weights | No | Yes | — |
| Release | Mar 1, 2023 | Aug 28, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | OpenAI | Tencent Cloud TokenHub | — |
| Modalities | text → text | text → text | — |
| Specs | |||
| Context window | 16K | 1M | B +984K |
| Max output | 4K | - | — |
| Parameters | — | 770B (49B act.) | — |
| Pricing | |||
| Input $/1M | $0.50 | $0.83 | A +$0.33 |
| Output $/1M | $1.50 | $2.50 | A +$1.00 |
| Blended $/1M (3∶1) | $0.75 | $1.25 | A +$0.50 |
| Speed | |||
| tok/s | 120 | - | — |
| TTFT (s) | 0.25 | - | — |
| Reasoning | |||
| MMLU-Pro | 46.2 | — | — |
| GPQA Diamond | 29.7 | 92.3 | B +62.6 pts |
| Humanity's Last Exam | — | — | — |
| AIME 2025 | — | — | — |
| MATH-500 | 44.1 | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | — | — |
| AA-LCR v1.1 | — | — | — |
| CritPt | — | — | — |
| MMMU-Pro | — | — | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | 65.7 | — |
| SWE-bench Multilingual | — | 82.9 | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | 85.4 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | — | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | 64.3 | — |
| WebDev Arena | — | 1,624 | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,225 | — | — |
GPT-3.5 Turbo
- Context
- 16K / 4K out
- Parameters
- —
- Price
- $0.50 / $1.50
- Speed
- 120 tok/s · 0.25s TTFT
- Modalities
- text → text
- License
- Proprietary
Classic ChatGPT-era budget model — still referenced for cheap classification and simple chat.
Hy4 Preview
- Context
- 1M
- Parameters
- 770B (49B act.)
- Price
- $0.83 / $2.50
- Speed
- —
- Modalities
- text → text
- License
- Apache 2.0
Tencent Hy Team's Aug 28, 2026 open-weight flagship (770B MoE / 49B active, Apache 2.0) with a 1M context window and default high reasoning effort. Text-only instruct weights on Hugging Face; hosted API at $0.834/$2.501 per 1M tokens via TokenHub and OpenRouter.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.