vust

AI MODELS IN TELEGRAM · 23 SUPPORTED

Pick the model by the task, not the brand

23 supported models in one Telegram chat — GPT-5.6, Claude, Gemini, Grok 4.5, Kimi K3, DeepSeek V4, Qwen 3.7 and more. Start free; Pro and Max can select the full catalog, and every run still re-checks availability, included limits and Balance.

Free tier to try. No card to start.23 current models
23 current modelsMatched by taskFree tier to try

The Short Version

Task first, model second

Quick work → an included fast model. Code → Claude Sonnet 5, DeepSeek V4 Pro or Kimi K3. Long documents → Gemini 3.1 Pro, Inkling or Kimi K3. High stakes → a flagship plus a cross-check with a second model.

@vustbot recommends a default, while All models keeps the full catalog one tap away.

Find your task

Models that usually fit each job, split by access. Full 23-model reference below.

Everyday & quick asks

Rewrites, quick answers, short plans. An included model is usually enough.

Included · Free/CoreGPT-5.6 LunaFREEGemini 3 FlashFREEDeepSeek V4 FlashFREEQwen 3.7 PlusCORELongCat 2.0COREGemini 3.6 FlashCOREGemini 3.5 Flash-LiteFREE
Flagships · ProGLM-5.2PRO

Flagships are rarely needed here — save the allowance for harder tasks.

Analysis & structured thinking

Summarize evidence, find contradictions, build a structured argument.

Included · Free/CoreGemini 3 FlashFREEGPT-5.6 TerraCOREClaude Sonnet 5COREDeepSeek V4 FlashFREEQwen 3.7 PlusCORELongCat 2.0COREGemini 3.6 FlashCOREGemini 3.5 Flash-LiteFREE
Flagships · ProGemini 3.1 ProPROGLM-5.2PROGrok 4.5PROKimi K3PROGPT-5.6 SolPRODeepSeek V4 ProPROMiniMax M3PROQwen 3.7 MaxPROInklingPRO

Coding & debugging

Multi-file refactors, tricky bugs, code review with reasoning.

Included · Free/CoreGPT-5.6 TerraCOREClaude Sonnet 5COREQwen 3.7 PlusCOREGemini 3.6 FlashCORE
Flagships · ProClaude Opus 4.8PROGLM-5.2PROKimi K3PROGPT-5.6 SolPRODeepSeek V4 ProPROMiniMax M3PROQwen 3.7 MaxPROClaude Fable 5PROFugu UltraPRO

Long documents & context

Reports, contracts, transcripts — work that has to hold a lot at once.

Included · Free/CoreQwen 3.7 PlusCORELongCat 2.0COREGemini 3.6 FlashCORE
Flagships · ProClaude Opus 4.8PROGemini 3.1 ProPROKimi K3PROMiniMax M3PROInklingPROFugu UltraPRO

Research

Comparing sources, synthesizing a position, structured deep dives.

Included · Free/CoreGemini 3.6 FlashCORE
Flagships · ProClaude Opus 4.8PROGemini 3.1 ProPROGrok 4.5PROGPT-5.6 SolPROQwen 3.7 MaxPROClaude Fable 5PROInklingPROFugu UltraPRO

High-stakes decisions

Work you will act on. Use a flagship — and cross-check with a second model.

Included · Free/CoreGPT-5.6 TerraCOREClaude Sonnet 5CORE
Flagships · ProClaude Opus 4.8PROGrok 4.5PROGPT-5.6 SolPRODeepSeek V4 ProPROClaude Fable 5PROFugu UltraPRO

Manual previews

Named entries you pick yourself — Auto does not select them yet.

Each has its own release page with verified facts, limits and a direct link.

What each plan includes

Free

The always-on everyday default plus a small daily allowance on the light models. Enough to try the product with no card.

Core

Everyday pre-flagships with larger included limits.

Pro & Max

The full 23-model catalog including flagships. Included allowances are bounded; every run still re-checks availability, limits and Balance.

Outside a plan or over an allowance, the bot shows the exact one-shot Balance option before anything runs — your prompt is kept, and nothing is charged before you confirm.

The full catalog

Every model is selected by its full name — the VUST label plus the canonical model name — and the bot confirms it before the run.

Free4

VUST Fast

GPT-5.6 Luna

Everyday questions

VUST Smart

Gemini 3 Flash

Everyday questions · Analysis

VUST Efficient

DeepSeek V4 Flash

Analysis · Everyday questions

VUST Gemini Lite

Gemini 3.5 Flash-Lite

Everyday questions · Analysis

Core5

VUST Deep

GPT-5.6 Terra

Analysis · Coding · High-stakes decisions

VUST Claude

Claude Sonnet 5

Coding · Analysis · High-stakes decisions

VUST Qwen Plus

Qwen 3.7 Plus

Everyday questions · Analysis · Coding · Long documents

/qwen

VUST LongCat

LongCat 2.0

Everyday questions · Analysis · Long documents

VUST Gemini

Gemini 3.6 Flash

Everyday questions · Analysis · Coding · Research · Long documents

Pro & Max14

VUST Max

Claude Opus 4.8

Coding · Research · High-stakes decisions · Long documents

VUST Long

Gemini 3.1 Pro

Research · Long documents · Analysis

VUST GLM

GLM-5.2

Analysis · Coding · Everyday questions

VUST Grok

Grok 4.5

Analysis · Research · High-stakes decisions

/grok

VUST Kimi

Kimi K3

Coding · Analysis · Long documents

/kimi

VUST GPT Flagship

GPT-5.6 Sol

High-stakes decisions · Coding · Analysis · Research

VUST DeepSeek Pro

DeepSeek V4 Pro

Analysis · Coding · High-stakes decisions

VUST MiniMax

MiniMax M3

Analysis · Coding · Long documents

/minimax

VUST Qwen Max

Qwen 3.7 Max

Analysis · Coding · Research

/qwen

VUST Fable

Claude Fable 5

Coding · Research · High-stakes decisions

VUST Muse

Muse Spark 1.1

Analysis · Coding · Research

/muse-spark

VUST Inkling

Inkling

Analysis · Research · Long documents

VUST Fugu

Fugu Ultra

Coding · Research · High-stakes decisions · Long documents

VUST Opus

Claude Opus 5

Coding · Research · High-stakes decisions · Long documents · Analysis

/claude-opus-5

Grouped by plan: Free, Core, then Pro & Max. Availability, included limits and Balance requirements are re-checked before every run.

Copyable summary

Citable summary (checked 2026-07-26): VUST supports 23 current AI models in one Telegram chat (@vustbot), grouped by task — everyday, analysis, coding, research, long context, high-stakes. Free includes light models; Pro and Max can select the full catalog, with Claude Opus 5 and Muse Spark 1.1 as manual previews. Every run names the model and re-checks access, limits and Balance.

Frequently asked questions

Do Pro and Max really unlock all 23 models?

Yes, all 23 supported models are visible and selectable in the native catalog. Included allowances depend on the plan; past the limit the bot shows the exact Balance price before it runs.

Do I need my own API keys?

No. You select the model inside Telegram, and VUST shows its familiar name before the run. No separate cloud console or manual endpoint setup is required.

Is there a separate "thinking mode" to switch on?

No — VUST does not have a distinct fast/thinking toggle. Picking a stronger named model over a fast one is what changes the depth of reasoning.

What happens when I hit a limit?

Your prompt is kept. The bot shows Pro/Max plans, the exact one-shot Balance price for this run, and an explicit switch back to an included model — nothing is charged before you confirm.