Up to date
tencent Other out 6 months ago
299B parameters (8 of 192 experts per token), text in, text out, 256K tokens of context. 1 provider serves it from $0.18 per million input tokens.
Verified 2026-10-25 Rank #684 in LLMs · score 72.4 method v1 Sources 4 Downloads 30d 81.6k · #336 of 619 tracked
Picked by rule from 6 tracked pieces: the lab's own post, the most shared, a field report, the latest.
Hy3 preview from Tencent tops OpenRouter rankings despite DeepSeek V4 Flash offering significantly lower effective pricing of $0.018/1M input tokens via a 2% cache read cost.
Tencent released Hy3 on July 6, 2026, as an updated version of its Hunyan model available under the Apache 2.0 license.
The 2 picks above, plus these 4. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.
Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.
| # | Cheapest live row | In /1M | Out /1M | Context | Quant | Seen |
|---|---|---|---|---|---|---|
| 1 | GMICloud | $0.18 | $0.60 | 256K | bf16 | Oct 24 |
Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
Served by 1 API provider; 8 quantized builds for local use.
Top Spaces running it: Tracy146/C-Hermes-Agent ♥ 8 · lablab-ai-amd-developer-hackathon/Threat_Hunter ♥ 1 · ICSAC-Institute/editorial-system ♥ 1 · victor/space-thumbnail-previews-emoji-top ♥ 1
Weights only, an estimate: leave room for the KV cache. Method
About 179 GB at INT4, 359 GB at FP8, 717 GB at BF16: weights only, so leave room for the KV cache. Every expert must be resident, so total parameters count.
Estimate params × bytes per param × 1.2 · how we estimate
Weights, every dated article, and repo commits on one rail.
Teal is the weights (2026-04-13), amber triangles are 6 dated articles (each links to it), grey ticks are repo commits.
11 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Everything else, folded: the full tables, every source.
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
| Pipeline tag | text-generation |
|---|---|
| Library | transformers |
| Architectures | HYV3ForCausalLM |
| Model type | hy_v3 |
| Params by dtype | BF16: 298,786,140,416 · F32: 15,360 |
| Size band | XXL |
| MoE | yes |
| Experts | 192 |
| Experts per token | 8 |
| Context length | 262,144 (config.json max_position_embeddings) |
| License | Other (other) |
| Hub snapshot | 2026-10-25 |
vLLM: supported · HYV3ForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: not in registry · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed
vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.
Weights-only memory for Hy3-preview: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 717 GB · Cluster (≥640 GB) |
|---|---|
| FP8 (1 byte/param) | 359 GB · Multi-GPU node (<640 GB) |
| INT4 (0.5 bytes/param) | 179 GB · Multi-GPU node (<640 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra. MoE: all experts must be resident, so total parameters count.
quantized: 8finetune: 2adapter: 1
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| mlx-community/Hy3-preview-4bit | quantized | -4bit | 295B | XXL | other | mlx-community | 4 | 3.3k |
| 0xSero/Hy3-299B-NVFP4 | quantized | NVFP4 | 169B | XXL | other | 0xSero | 5 | 2.2k |
| JANGQ-AI/Hy3-preview-JANGTQ_K | quantized | — | 36.5B | XL | other | JANGQ-AI | 3 | 1k |
| mlx-community/Hy3-preview-8bit | quantized | -8bit | 295B | XXL | other | mlx-community | 1 | 1k |
| mlx-community/Hy3-preview-6bit | quantized | — | 295B | XXL | other | mlx-community | 1 | 865 |
| JANGQ-AI/Hy3-preview-JANGTQ | quantized | — | 30.5B | L | other | JANGQ-AI | 0 | 570 |
| 0xSero/Hy3-299B-FP8 | quantized | -FP8 | 299B | XXL | other | 0xSero | 0 | 512 |
| LibertAIDAI/Hy3-preview-NVFP4 | quantized | NVFP4 | 172B | XXL | other | LibertAIDAI | 0 | 300 |
| spellbound-eng/hy3-spellbound-50k-merged-sft-step2677 | finetune | — | 295B | XXL | other | spellbound-eng | 0 | 3 |
| spellbound-eng/hy3-spellbound-simpo-r256-b10-g3-sft0p1-lr2e-6-step151-lora | adapter | — | — | — | other | spellbound-eng | 0 | 0 |
| mlx-community/Hy3-preview-MTP-4bit | finetune | -4bit | — | — | other | mlx-community | 0 | 0 |
| Date | Commit | SHA |
|---|---|---|
| 2026-04-23 | update readme. (#4) | 549c2b3 |
| 2026-04-23 | Update README.md | bda30e4 |
| 2026-04-23 | Upload folder using huggingface_hub | bda4c91 |
| 2026-04-23 | Update README.md | ab24fda |
| 2026-04-23 | Upload folder using huggingface_hub | cf1003d |
| 2026-04-22 | Upload folder using huggingface_hub | f332661 |
| 2026-04-22 | Upload folder using huggingface_hub | b53bd70 |
| 2026-04-22 | Add files using upload-large-folder tool | 0de7283 |
| 2026-04-22 | Add files using upload-large-folder tool | 166e27d |
| 2026-04-22 | Add files using upload-large-folder tool | 1a01fbf |
| 2026-04-13 | initial commit | a4cac00 |
Hosted endpoints observed for Hy3-preview, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.18/M.
| Provider | Model ref | Context | Quant | $/M in | $/M out | tok/s | TTFT | Tools | Status | Observed |
|---|---|---|---|---|---|---|---|---|---|---|
| GMICloud openrouter | tencent/hy3-preview | 256K | bf16 | $0.18 | $0.60 | — | — | yes | live | 2026-10-24 |
Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:42 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:13 UTC · Last verified 2026-10-25
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.
