Hy3 vs DeepSeek-V4-Flash-0731-NVFP4

Specs, VRAM requirements and download trends — updated 10 November 2026.

tencent · A · Open weights
Params
298.8 B
Context
262,144
30d
1.2 M
apache-2.0 · commercial OK
MJPansa · B · Open weights
Params
304.2 B
Context
1,048,576
30d
203 K
mit · commercial OK

Specification comparison

Differences are highlighted; identical values are muted.

Specification Hy3 DeepSeek-V4-Flash-0731-NVFP4
Deployment Open weights · also via API Open weights
Parameters 298.8 B 304.2 B
Architecture HYV3ForCausalLM DeepseekV4ForCausalLM
Context length 262,144 1,048,576
Licence apache-2.0 mit
Commercial use Allowed Allowed
Languages — —
Downloads 30d 1,161,460 202,545
Downloads all time 1.7 M 460 K
Quantizations on the Hub SAFETENSORS SAFETENSORS
Gated No No
First seen on the Hub 2026-07-02 2026-07-31
API input price / 1M tokens $0.13 not on OpenRouter
API output price / 1M tokens $0.53 not on OpenRouter
Est. cost per 1M input + 1M output tokens $0.66 not on OpenRouter
Cached input / 1M tokens $0.033 not on OpenRouter
AA Intelligence Index — not on OpenRouter
AA Coding Index — not on OpenRouter
AA Agentic Index — not on OpenRouter
Max output tokens 128,000 not on OpenRouter
API providers 5 not on OpenRouter
Tool calling Yes not on OpenRouter
Reasoning mode Yes not on OpenRouter

API cost

Estimated cost of 1M input + 1M output tokens via OpenRouter

Hy3 $0.66

Download trend

Daily snapshots, last 43 days (29 Sep – 10 Nov)

1.4 M175 K
Hy3 DeepSeek-V4-Flash-0731-NVFP4 A overtook B on 3 Oct

VRAM side by side

On RTX 4090 · 24 GB · 8K context unless noted

Hy3 · fp16 @ 256K ctx 2,092.0 GB / 24 GB
❌ Won’t fit
DeepSeek-V4-Flash-0731-NVFP4 · fp16 @ 1024K ctx 6,033.9 GB / 24 GB
❌ Won’t fit

Adjacent comparisons