Hy3
Hy3Alternatives?
CheaperMiMo v2.6 FlashunstatedOpenwins 54%-24%$0.14 / $0.28Memory (4-bit): 179 GB → 186 GB+4
BetterGPT-5.6 Terraxhighwins 54%+1,848%$0.63Prompts with >272K input tokens
OpenGLM 5.3 Flashunstatedwins 56%same$0.15 / $0.50Memory (4-bit): 179 GB → 193 GB+5
CheaperGemma 4 31BunstatedOpensame-31%$0.10 / $0.34CheaperMiMo v2.5unstatedOpenwins 47%-24%$0.14 / $0.28Memory (4-bit): 179 GB → 186 GBMiMo v2.6 FlashunstatedOpenwins 54%-24%$0.14 / $0.28Memory (4-bit): 179 GB → 186 GB+4Gemma 4 31BunstatedOpensame-31%$0.10 / $0.34MiMo v2.5unstatedOpenwins 47%-24%$0.14 / $0.28Memory (4-bit): 179 GB → 186 GBMiMo v2 Omniunstatedwins 47%-24%$0.14 / $0.28
Qwen3 235B A22B Instruct 2507unstatedOpenwins 47%-11%$0.09 / $0.55GPT-5.6 Terraxhighwins 54%+1,848%$0.63Prompts with >272K input tokens
Muse Spark 1.2xhighwins 53%+766%$0.97Only auto tool_choice is supported+1GPT-5.6 Solxhighwins 56%+3,363%$1.18Gemini 3.8 Flashhighwins 59%+549%*$1.24Sampling parameters are ignored+1GLM 5.2maxOpenwins 55%+831%$1.47Memory (4-bit): 179 GB → 452 GBGLM 5.3 FlashunstatedOpenwins 56%same$0.15 / $0.50Memory (4-bit): 179 GB → 193 GB+5MiMo v2.6 FlashunstatedOpenwins 54%-24%$0.14 / $0.28Memory (4-bit): 179 GB → 186 GB+4
Step 5 PreviewunstatedOpenwins 53%+517%$1 / $2.70Memory (4-bit): 179 GB → 363 GB+3Qwen3.8 27BunstatedOpensame+314%$0.42 / $2.55max_tokens range · QwenCloud+4Qwen3.5 397B A17BunstatedOpensame+484%$0.60 / $3.60Memory (4-bit): 179 GB → 242 GBRun it yourself
estimateBenchmarks
in Chat & KnowledgeFAQ
How much does Hy3 cost?›
$0.13 per 1M input tokens and $0.53 per 1M output tokens; cached input $0.033 per 1M; cheapest via DeepInfra at $0.13 / $0.53. A typical benchmark task costs about $0.072 at default effort.
Is Hy3 free to use?›
There is no free API; the cheapest paid price is $0.13 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Hy3, and what is it best at?›
#71 of 334 in Chat & Knowledge, #63 of 132 in Coding, #63 of 133 in Agents & Tool Use, #52 of 195 in Writing & Language, #52 of 133 in Vision & Documents. Strongest at: Reasoning, Science, Long context.
Is Hy3 better than Qwen3.6 Max Preview?›
On Arena Text · Chinese, Hy3 scores 1,487.5 vs 1,488 — about the same.
How can I use Hy3?›
Through the API of AtlasCloud, DeepInfra, Featherless, GMICloud, Novita and 2 more.
What is the context window of Hy3?›
256K tokens, with up to 128K output tokens.
When was Hy3 released? Is there a newer version?›
Released 2026-07-06. It is the current version in its line.
Is Hy3 open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Hy3 locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Hy3 read images?›
No, it accepts text only.
What is Hy3?›
Hy3 is an open-weights language model from Tencent, released Jul 2026, the successor to Hy3 preview.
What are the best alternatives to Hy3?›
MiMo v2.6 Flash (cheaper), GPT-5.6 Terra (better), GLM 5.3 Flash (open weights).