vs
4:0
overall capability wins (0 ties across 4 scored). Hy4 Preview leads overall.
- Reasoning1:0
- Coding2:0
- Pricing0:3
- Specs1:0
Verdict
Hy4 Preview leads coding (widest gap: +457.7 on WebDev Elo); Hy4 Preview offers ~7.8x larger context (1M vs 128K); Hy4 Preview ships open weights (Apache 2.0).
- Hy4 Preview leads coding (widest gap: +457.7 on WebDev Elo)
- Hy4 Preview offers ~7.8x larger context (1M vs 128K)
- Hy4 Preview ships open weights (Apache 2.0)
- Mercury 2 is ~3.3x cheaper on a blended token basis than Hy4 Preview
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Hy4 Preview
- $1.46
- Mercury 2
- $0.44
Mercury 2 is about 3.3x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Tencent
Inception Labs
Family
Other
Other
License
Apache 2.0
Proprietary
Open weights
Yes
No
Release
Aug 28, 2026
Sep 1, 2026
Knowledge cutoff
-
-
API / provider
Tencent Cloud TokenHub
Inception Labs
Modalities
text → text
text → text
Specs
Context window
1M
128K
Max output
-
-
Parameters
770B (49B act.)
—
Pricing
Input $/1M
$0.83
$0.25
Output $/1M
$2.50
$0.75
Blended $/1M (3∶1)
$1.25
$0.38
Speed
tok/s
-
-
TTFT (s)
-
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
92.3
77.0
Humanity's Last Exam
—
17.1
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
21.2
AA-LCR v1.1
—
43.7
CritPt
—
0.8
MMMU-Pro
—
—
IFBench
—
69.8
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
27.3
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
37.7
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
64.3
—
WebDev Arena
1,624
1,167
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
—
| Metric | Hy4 Preview | Mercury 2 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Tencent | Inception Labs | — |
| Family | Other | Other | — |
| License | Apache 2.0 | Proprietary | — |
| Open weights | Yes | No | — |
| Release | Aug 28, 2026 | Sep 1, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Tencent Cloud TokenHub | Inception Labs | — |
| Modalities | text → text | text → text | — |
| Specs | |||
| Context window | 1M | 128K | A +872K |
| Max output | - | - | — |
| Parameters | 770B (49B act.) | — | — |
| Pricing | |||
| Input $/1M | $0.83 | $0.25 | B +$0.58 |
| Output $/1M | $2.50 | $0.75 | B +$1.75 |
| Blended $/1M (3∶1) | $1.25 | $0.38 | B +$0.88 |
| Speed | |||
| tok/s | - | - | — |
| TTFT (s) | - | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 92.3 | 77.0 | A +15.3 pts |
| Humanity's Last Exam | — | 17.1 | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 21.2 | — |
| AA-LCR v1.1 | — | 43.7 | — |
| CritPt | — | 0.8 | — |
| MMMU-Pro | — | — | — |
| IFBench | — | 69.8 | — |
| 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 | 27.3 | A +58.1 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | 37.7 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | 64.3 | — | — |
| WebDev Arena | 1,624 | 1,167 | A +458 Elo |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | — | — |
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.
Mercury 2
- Context
- 128K
- Parameters
- —
- Price
- $0.25 / $0.75
- Speed
- —
- Modalities
- text → text
- License
- Proprietary
Inception Labs' diffusion-based (non-autoregressive) reasoning LLM, API-only with a 128K context window and high-throughput inference on Blackwell GPUs.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.