ONETEN · WORKFLOW ATLAS · 2026.07

ENGINE ROOM

A visual atlas of every ONETEN automation pipeline. See at a glance which LLMs and tools cooperate inside each workflow.

ANTHROPIC·DEEPSEEK·OPENAI·GOOGLE·MINIMAX·FAL·HIGGSFIELD
§ 01 · MODEL GALLERY

Model Gallery

Each pedestal is one LLM or tool inside the ONETEN stack — a real 3D exhibit you can spin. Beneath it: the pipelines it currently powers.

AI
Claude Fable 5 · Opus 4.8
Anthropic

Primary interaction · agent orchestration · code refinement. Fable 5 (Mythos-class) is the session default since July — survived a 18-day US export-control freeze; Opus 4.8 stays as the flat-rate fallback.

· Max plan · Fable 5 draws usage credits after 7/7 ·
Claude Code · interactive buildds CLI · heavy-lift offload from Claude
CX
Codex
OpenAI

OpenAI's coding agent — and the CC storyboard engine's realtime image generator: view_image anchors + built-in image_gen produce face-locked cinematic shots per beat, zero API fees on subscription.

· subscription · image_gen free ·
Claude Code · interactive buildCC · realtime storyboard engine
AG
Antigravity 2.0
Google

Google's agentic coding IDE (latest 2.0). Another parallel-agent surface in the dev workflow alongside Claude Code and Codex.

· newly integrated ·
Claude Code · interactive build
DS
DeepSeek V4
DeepSeek

Heavy-lift outsourcing: long-context analysis, dialogue writing, boilerplate, refactor planning. Cheap.

· ≈ 1/10 of Claude ·
Haifan v2 · ChineseHaifan v2 · EnglishHaifan v2 · Koreands CLI · heavy-lift offload from ClaudeBuild log · auto-publishPA1 · 9:16 vertical commentaryMaster-call v3 · MCV3 with model rotationCC · realtime storyboard engineONE TEN ORACLE · multi-model vote
G
GPT-5.6 · Sol / Terra / Luna
OpenAI · 2026-07-09

OpenAI's new three-tier family (Sol = frontier reasoning · Terra = 2x cheaper than 5.5 · Luna = fast/cheap). Legacy GPT-5.5 handled Haifan v1 storyboards; 5.6 tiers are the new option pool for ORACLE votes and dialogue rotation.

· Sol $5/$30 · Terra $2.5/$15 · Luna $1/$6 per M ·
Haifan v1 · storyboards (legacy)image_genONE TEN ORACLE · multi-model vote
GL
GPT-Live
OpenAI · 2026-07-08

Full-duplex voice model — listens while it speaks, real-time interpretation, delegates complex asks to GPT-5.x. API pending; earmarked as the future voice layer of the ONE TEN living room.

· ChatGPT rollout · API waitlist ·
G4
Gemma 4
Google · local Ollama

Local resident: OCR, translation, polish, default LLM for the dali bot. Free batch runs.

· Local · free ·
dali_botBuild log · auto-publishQian-Yi-Shi · feeling journal3in · cross-platform broadcastONETEN Company · meeting videos
MM
MiniMax Hailuo
MiniMax

TTS — 4-character bilingual voice synthesis (speech-02-hd). Since July routed through the Fal relay (same model + voice ids, pay-per-use pennies) after MiniMax raised its direct recharge minimum to $25. All calls pass tts_normalize first.

· via Fal · $0.10 / 1k chars ·
Haifan v2 · ChineseHaifan v2 · EnglishFujimasa · character lip-syncONETEN Company · meeting videosPA1 · 9:16 vertical commentaryMaster-call v3 · MCV3 with model rotationTTS Normalizer · multi-language read-fixCC · realtime storyboard engine
HF
Higgsfield
Higgsfield AI

Lip sync + image gen (gpt_image_2 / Seedream / Flux). Lip sync for character portraits, gpt_image_2 for hero infographics, Seedream for ref-based avatar pool.

Fujimasa · character lip-syncOmni · motion graphic heroPA1 · 9:16 vertical commentary
HX
HyperFrames
HeyGen open-source

HTML → MP4 render layer: composes storyboards + dialogue + voice into 1080p output.

Haifan v2 · ChineseHaifan v2 · EnglishHaifan v2 · KoreanONETEN Company · meeting videosPA1 · 9:16 vertical commentaryMaster-call v3 · MCV3 with model rotation
GF
Gemini 3.5 Flash
Google · I/O 2026-05-19

Frontier-tier reasoning at half-price. Drop-in upgrade for fast LLM calls when DeepSeek is too slow.

· ≈ 1/2 of Claude Sonnet ·
Master-call v3 · MCV3 with model rotation
GO
Gemini Omni
Google DeepMind

Multimodal video gen — text/image/audio/video in, physics-aware video out. Used for PA1 motion graphic hero scenes (SPY chart, McKinsey infographic).

· ≈ $0.20-0.60/sec output ·
PA1 · 9:16 vertical commentaryOmni · motion graphic hero
OR
OpenRouter
OpenRouter aggregator

Aggregator: 9-model rotation for MCV3 dialogue + the ORACLE voting panel (GLM-5.2 / Grok-4.3 / Gemini-3.5 / GPT-5.4-mini and friends). Model ids validated against /v1/models before every run.

· DeepSeek-tier pricing across all ·
Master-call v3 · MCV3 with model rotationONE TEN ORACLE · multi-model vote
§ 02 · PIPELINE ATLAS

Pipeline Atlas

One card per pipeline. Click a card to expand the flow + description.