Skip to main content
BenchLM
Data

Data as of November 19, 2026 · How the score is built

Superseded.Tencent has newer models in this line:Hy4 preview

Hy3

SupersededOpen WeightReasoning

Decision readingHy3 is tracked, but not publicly ranked yet. The profile exposes 3 sourced benchmark rows and leaves unsupported fields blank until a published record exists.

Email me when its price, rank or availability changes.

Capability

Unranked

field median 49Not eligible for a public rank

Price

Self-hosted; infrastructure cost varies

input median $0.60No comparable first-party hosted token rate

Speed

77tok/s

field median 93 tok/sFirst token 41.1 s

Context

256Ktokens

field median 256,000Reported for this model; direct source link not stored

Decision snapshot

How much published evidence this model has, where that evidence is strongest, and what to validate before choosing it. It has no public rank, so no field position is shown.

Strongest published evidence

Coding ranks #53. Particularly well-suited for software development and code generation tasks.

Validate before choosing

3 published rows leave some tracked benchmark slots empty. No comparable first-party API token rate is published.

Source-linked · 3 displayable benchmark rows

Is Hy3 right for your job?

Its place on the LLM Selector’s job rankings, the same order the job pages publish. Describe your own job to see where it stands for that.

Describe the task and constraints. This selector recommends models; it cannot carry out the task or access your files and accounts.

We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy

Category score record

Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #89 of 138Percentile 36thWeight 22%Not listedUnlisted sources
35.6
CodingRank #53 of 160Percentile 67thWeight 20%Not listedUnlisted sources
46.4
ReasoningWeight 17%0 benchmarksNot measured
Not measured
MultimodalWeight 12%0 benchmarksNot measured
Not measured
KnowledgeWeight 12%0 benchmarksNot measured
Not measured
MultilingualWeight 7%0 benchmarksNot measured
Not measured
Inst. FollowingWeight 5%0 benchmarksNot measured
Not measured
MathWeight 5%0 benchmarksNot measured
Not measured

3 of 923 tracked benchmark slots have displayable evidence · bars run 0–100

Coverage details

How much of this is verified

Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.

  1. AgenticNot measured
  2. CodingNot measured
  3. ReasoningNot measured
  4. MultimodalNot measured
  5. KnowledgeNot measured
  6. MultilingualNot measured
  7. Inst. FollowingNot measured
  8. MathNot measured
Verified sourceProvisionalNot measured

Benchmark ledger

Every result says who ran the test. External signals opens by default. The tick marks the best source-verified result in the catalog; provisional leaders do not set the reference.

Overall rank
Not publicly ranked
Lab-reported
0
None
Run by a benchmark
3
The benchmark’s own run
Independent runs
0
None yet
External signals3 rows
External signals benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
SkillsBench official leaderboard (BenchFlow)SkillsBench v1.1 official leaderboard (BenchFlow)55.9%

Best verified: GPT-5.5 · 67.3%

Gap11.4 behindRun by the benchmarkShown only
NYT ConnectionsNYT Connections LLM Benchmark (extended version)41.2%

Best verified: Gemini 4 Carbon · 99.3%

Gap58.1 behindRun by NYT ConnectionsShown only
BabaIsBenchBabaIsBench (Baba Is You intro levels, pass@1)54%

Best verified: Claude Opus 5 · 100%

Gap46 behindRun by BabaIsBenchShown only

Every bar is benchmark-run or independent. Bars run 0–100; the dark tick marks the best source-verified value.

Lineage

The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.

  1. Apr 23, 2026

    Hy3 Preview

    Not publicly ranked · Price not listed

  2. Jul 6, 2026 · you are here

    Hy3

    Not publicly ranked · Price not listed

  3. Aug 28, 2026

    Hy4 preview

    Score 60.4 · Price not listed

Instruct

Hy3 Preview

Spec sheet

Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.

API model ID
Not published
Context window
256K
Maximum output
Not sourced yet
Knowledge cutoff
Not sourced yet
Input modalities
Not sourced yet
Output modalities
Not sourced yet
Parameters
Not sourced yet
Availability
Not sourced yet
Cloud regions
Not tracked yet
Lifecycle
Superseded
API capabilities
Tool calling, structured outputs, and batch support are not tracked yet
Prompt caching
Not documented in the pricing record
Self-host
Hy3 needs ~320GB VRAM (4× NVIDIA H100 (80GB)).
Rate limits
Not tracked yet

How to read this profile

The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.

We track Hy3, but the public leaderboard excludes this profile until enough non-generated benchmark coverage is available. Only published rows appear above.

Hy3 is a open weight model with a 256K context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.

Tracked from Tencent's tencent/Hy3 Hugging Face model card. Tencent describes Hy3 as a 295B-total-parameter MoE with 21B active parameters, 3.8B MTP layer parameters, Apache-2.0 weights, and 256K context. The model card's benchmark appendix is rendered as an image in the reviewed public page, so BenchLM keeps scored benchmark fields empty until exact text-published values are available.

Hy3 sits in the Hy3 family with Hy3 Preview. Its explicit predecessor is Hy3 Preview. 3 of 923 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.

Its strongest eligible category is Coding at #53, while its lowest eligible position is Agentic at #89. particularly well-suited for software development and code generation tasks.

Last updated November 19, 2026. Runtime fields remain blank until a sourced snapshot exists.

Deployment options

Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.

Where to run Hy3 by API

2 hosts publish a per-token price for it. OpenRouter has the lowest output rate, $0.53 per million tokens. Prices read from each host’s own pages on November 16, 2026.

Inference hosts for Hy3 with input, output and cached-input prices per million tokens.
HostInput /1MOutput /1MCached /1MContext
OpenRouter$0.13$0.53$0.033262K
Novita AI$0.14$0.58$0.035262K
Compare inference hosts

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Hy3
API / mo (Novita AI)$540
Self-host / mo$13,981
Break-evenNot reached at 70% load
Model the full break-even

Questions

How does Hy3 perform overall in AI benchmarks?

Hy3 has 3 source-displayable benchmark rows, but it does not qualify for a public overall rank. The available rows remain visible by category without being converted into a site-wide score. Missing evidence stays blank instead of being estimated from an earlier model.

Is Hy3 open source?

Hy3 is an open-weight model from Tencent. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.

Which sibling models are related to Hy3?

Hy3 belongs to the Hy3 family. Related tracked variants include Hy3 Preview. A sibling link indicates shared lineage or a documented configuration relationship; it does not mean the variants have identical pricing, context limits, benchmark evidence, or deployment behavior. Compare before switching.

Does Hy3 have full benchmark coverage on BenchLM?

No. Hy3 currently has 13 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.

What is the context window size of Hy3?

Hy3 has a reported context window of 256K in the exact-model catalog record. The value stays visible, but the profile marks its source link as unavailable instead of presenting it as directly documented. Maximum output length remains separate because providers often publish a different limit.

Compare Hy3 with every tracked model1032 comparisons