{"id":"2cf3a1a6-06b9-4123-b93f-95eb12acb5b4","companyId":"a2532e93-8f76-49ef-b384-73186fff5021","modelName":"Heron / DriveHeron","slug":"turing-heron","description":"Turing's multimodal AI foundation model for autonomous driving. DriveHeron is a Vision-Language-Action (VLA) model with ~2 billion parameters that integrates camera data with language-based contextual understanding to predict and execute steering, acceleration, and braking. Achieved real-time autonomous driving on public roads at 10 Hz in March 2026, a first in Japan.","formFactor":"av","industryVertical":null,"environment":null,"maturityStage":"pilot","lifecycleState":"active","supersededByModelId":null,"specs":{},"manufacturerSerial":null,"reviewStatus":"reviewed","sources":[{"url":"https://tur.ing/en/news/20260326/","label":"Turing Achieves Japan's First Real-Time Autonomous Driving Using VLA Model","title":"Turing Achieves Japan's First Real-Time Autonomous Driving Using VLA Model","mediaType":"article","sourceName":"Turing Inc. (official)"}],"aliases":[],"collisionRisk":"low","reviewNote":null,"manufacturerTermForTeleop":null,"keyFacts":[{"label":"Form Factor","value":"av"},{"label":"Maturity","value":"pilot"},{"label":"VLA size and control rate","value":"Turing said on 26 March 2026 that its on-vehicle VLA has about 2 billion parameters and ran real-time inference and control at 10 Hz on public roads, which it called a Japan first based on its own March 2026 research ([Turing](https://tur.ing/en/news/20260326/))."},{"label":"RACER and DriveTiTok releases","value":"The same post says Turing released the RACER causal-reasoning driving dataset and DriveTiTok, an image tokenizer that compresses driving video to about one-hundredth of its size, with a portion of the data and models on Hugging Face ([Turing](https://tur.ing/en/news/20260326/))."}],"createdAt":"2026-08-20T07:04:16.167Z","updatedAt":"2026-10-03T17:36:47.178Z","unitsInFieldClaimed":null,"unitsInFieldVerified":null,"unitsManufacturedClaimed":null,"unitsManufacturedVerified":null,"unitsAsOf":null,"unitsSource":null,"salesModel":null,"priceUndisclosedSince":null,"autonomyClass":null,"autonomyStatement":null,"autonomyStatementSource":null,"demoForensics":null,"exportControlClass":null,"minCommitment":null,"warrantyTerms":null,"serviceModel":null,"productionRateClaimed":null,"productionRateEvidenced":null,"teleopRatioDisclosed":null,"interventionRateDisclosed":null,"jsonLd":{"@context":"https://schema.org","@type":"Product","@id":"https://registry.deploy.report/models/turing-heron","url":"https://registry.deploy.report/models/turing-heron","name":"Heron / DriveHeron","description":"Turing's multimodal AI foundation model for autonomous driving. DriveHeron is a Vision-Language-Action (VLA) model with ~2 billion parameters that integrates camera data with language-based contextual understanding to predict and execute steering, acceleration, and braking. Achieved real-time autonomous driving on public roads at 10 Hz in March 2026, a first in Japan.","identifier":"2cf3a1a6-06b9-4123-b93f-95eb12acb5b4","category":"av","publisher":{"@id":"https://deploy.report/#organization"}},"framework_metadata":{"framework_schema_version":"0.1.0","verification_status":"verified","maturity_stage":"pilot","lifecycle_state":"active","architectural_position":{"cohort":"av","sub_cohorts":[]},"within_cohort_verified_vs_claimed_pair":null,"cap_flags":[],"verification_depth":{"sources_count":1,"primary_source_types":["primary-company-ir"]}}}