Tencent
Hy4 Preview
Hy4 Preview is a model with open weights from Tencent. On the Vals AI Legal Research Bench it reaches 45.19 percent all-pass, position 21 of 91, within a standard error of ±3.46 percentage points at $0.85 per task, as at 9 November 2026 [1]. It does not pass the compliance gate: Open weights, but no route found at Vertex AI (EU), Azure Foundry (Data Zone EU), AWS Bedrock (EU regions) or the Mistral EU endpoint: the gate only opens on a dedicated EU endpoint or self-hosting, where the region is fixed by contract (as in the existing open-weights assessment).
- Public preview
- Open weights
- 28 August 2026
- PrudentScore
- -No score: the gate is not open
- Compliance gate
- RejectedOpen weights, but no route found at Vertex AI (EU), Azure Foundry (Data Zone EU), AWS Bedrock (EU regions) or the Mistral EU endpoint: the gate only opens on a dedicated EU endpoint or self-hosting, where the region is fixed by contract (as in the existing open-weights assessment)
- Cost per task
- $0.85Vals AI, per completed task [1]
01
Compliance gate
Why
- Open weights, but no route found at Vertex AI (EU), Azure Foundry (Data Zone EU), AWS Bedrock (EU regions) or the Mistral EU endpoint: the gate only opens on a dedicated EU endpoint or self-hosting, where the region is fixed by contract (as in the existing open-weights assessment).
The gate only opens when all three requirements are covered: processing in the EU, a contracting party in the EU or EEA, and a contractual exclusion of training on customer data. The gate tests a route, not a model name: the same model can pass through one provider and fail through another. Where a provider publishes no figure, we give no estimate.
Sources for this verdict
- docs.cloud.google.com/gemini-enterprise-agent-platform/resources/data-residency
- learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure-region-availability
- docs.aws.amazon.com/bedrock/latest/userguide/models-region-compatibility.html
- docs.mistral.ai/inference/regional-inference
02
Routes and EU processing
The gate tests a route (processor, region and contract), not a model name. Per route: processing in the EU, the contracting party and training on customer data.
No route has been recorded in the register for this model yet.
03
Scores per benchmark
Every measurement with its source, measurement date and the day we read it. The position counts only the models measured on the same benchmark and version.
- Legal Research BenchVersion 1 · Counts towards the PrudentScore45.19%± 3.4621 of 91
Vals AI, Legal Research Bench, v1, 9 November 2026 · Source · Measured 9 November 2026 · Read 10 November 2026 · cost per task $0.85
- Harvey LAB (Vals implementation)Version 1.18.75%± 2.1924 of 36
Vals AI, Harvey LAB (Vals implementation), v1.1, 9 November 2026 · Source · Measured 9 November 2026 · Read 10 November 2026 · cost per task $0.98
- Harvey LAB (Vals implementation)Version 1 · Counts towards the PrudentScore9.17%± 2.0022 of 80
Vals AI, Harvey LAB (Vals implementation), v1, 18 October 2026 · Source · Measured 18 October 2026 · Read 10 November 2026 · cost per task $0.98
- LegalBench (Vals leaderboard)Version 183.76%± 0.4148 of 113
Vals AI, LegalBench (Vals leaderboard), v1, 13 October 2026 · Source · Measured 13 October 2026 · Read 10 November 2026
- CaseLaw v2Not measured
- TaxEval v2Not measured
- Tax Agent BenchVersion 132.62%± 3.3821 of 82
Vals AI, Tax Agent Bench, v1, 9 November 2026 · Source · Measured 9 November 2026 · Read 10 November 2026 · cost per task $0.59
- CorpFin v2Not measured
- Immigration BenchNot measured
- Access to JusticeNot measured
- Public Benefits BenchVersion 1.168.61%± 1.219 of 50
Vals AI, Public Benefits Bench, v1.1, 12 October 2026 · Source · Measured 12 October 2026 · Read 10 November 2026 · cost per task $0.25
- Vals Legal IndexVersion 171.70%± 1.8413 of 36
Vals AI, Vals Legal Index, v1, 9 November 2026 · Source · Measured 9 November 2026 · Read 10 November 2026 · cost per task $0.81
- Vals IndexVersion 2.253.21%± 1.1325 of 35
Vals AI, Vals Index, v2.2, 9 November 2026 · Source · Measured 9 November 2026 · Read 10 November 2026 · cost per task $1.57
- Vals IndexVersion 2.149.94%± 1.1532 of 50
Vals AI, Vals Index, v2.1, 18 October 2026 · Source · Measured 18 October 2026 · Read 10 November 2026 · cost per task $1.44
- GPQA Diamond (Vals-run)Not measured
- MMLU Pro (Vals-run)Not measured
- Artificial Analysis Intelligence IndexNot measured
Source: Vals AI (Legal Research Bench, Harvey LAB (Vals implementation), LegalBench (Vals leaderboard), Tax Agent Bench, …) · Licence: CC BY 4.0 · Adaptation: Prudai selected the models, translated the explanations, added the gate verdict and the PrudentScore and rounded figures to two decimals.
The PrudentScore is a composition by Prudai B.V. based on publicly published figures and does not represent the view or endorsement of Vals AI, Harvey or Artificial Analysis. Figures from Artificial Analysis appear in running text only, with attribution and a link, as that party's terms require [9].
04
Key facts
- Developer
- Tencent
- Model family
- Hy
- Release date
- 28 August 2026
- Status
- Public preview
- Context window
- 1M tokens
- Max output tokens
- 64K tokens
- List price
- $0.83 input · $2.50 outputper million tokens · Read 10 November 2026
- Weights
- Open
- = not yet recorded in the register.
05
Frequently asked questions about this model
Does Hy4 Preview pass the compliance gate for legal work?
No. Open weights, but no route found at Vertex AI (EU), Azure Foundry (Data Zone EU), AWS Bedrock (EU regions) or the Mistral EU endpoint: the gate only opens on a dedicated EU endpoint or self-hosting, where the region is fixed by contract (as in the existing open-weights assessment). The gate tests a route, not a model name; the sources are on this page.
What does Hy4 Preview cost per task?
Vals computes $0.85 per completed task on the Legal Research Bench [1]. The list price is $0.83 per million input tokens and $2.50 per million output tokens.
How does Hy4 Preview score on legal research?
45.19 percent all-pass on the Vals AI Legal Research Bench, position 21 of 91 measured models, as at 9 November 2026 [1]. That benchmark measures US law; an independent, published benchmark for Dutch legal work does not exist yet.
Related models
Claude and Anthropic, GPT and OpenAI, Gemini and Google, Grok and xAI, Kimi and Moonshot AI, Muse and Meta, Mistral, DeepSeek, GLM and Z.ai, Harvey, Vals AI and Artificial Analysis are trademarks of their respective owners. They are named for identification only and this implies no partnership, sponsorship or endorsement.
Brand marks appear for identification only and imply no partnership, sponsorship or endorsement.