Subscribers
26 400
YOUTUBE SHOWS "26.4K subscribers"
CHANNELS / AI PROGBR
YouTube channel // numbers taken SEP 06, 2026
AI ProgBr
AI ProgBr has 26 400 YouTube subscribers and 1 540 500 views across the 91 videos we track, counted Sep 6, 2026.
Aqui você encontra tudo sobre modelos de IA. Fazemos testes, tutoriais e comparativos de ferramentas de inteligência artificial com foco em uso real: local, offline, open source e integradas a projetos.
▸ Verdict
0.64x
Views per video run at 64% of subscriber count — reach lags the subscriber base.
BASED ON 91 COLLECTED VIDEOS
Views · tracked
1 540 500
SUM OVER 91 VIDEOS WE FOLLOW — NOT LIFETIME
Avg views / video
16 928
OVER 91 TRACKED VIDEOS
Subscriber dynamics
FIRST MEASUREMENT
26 400
SEP 05, 2026
DAILY TRACKING BEGINS ON THE NEXT CRAWL.
GROWTH CHART APPEARS AT 2+ POINTS.
GROWTH CHART APPEARS AT 2+ POINTS.
Views across last 40 videos
Videos · 91
Gemini 3.8 Flash: the BEST Google model for everyday use?
GPT-6 Astra has SURPASSED humans. Has AGI finally ARRIVED?
Genjutsu: The most POWERFUL video model I've ever seen
Tencent Hy4: Another Chinese model on par with the GIANTS
I ran the new Qwen3.8 Flash Next LOCALLY: 125B at 35 t/s
Apple DESTROYS NVIDIA with the Mac Studio M5 Ultra. Is it for real?
Qwen 3.8 Flash Next: I tested the new architecture of the future "Qwen 4"
DeepSeek V4 Flash Vision: Native Vision vs. Sub-agent with Qwen 3.7 Plus. Who WINS?
Ox Alpha: the SECRET model that is SCARING people with its coding capabilities
The BEST model to run on 16GB VRAM: Qwen3.8 27B Q2 and Q3 are INSANE
Intel Arc Pro B70 32GB: 73 t/s with Qwen3.8-27B for 1/4 the price of an RTX 5090
I Turned Myself Into a 16-Bit Game Using Higgsfield MCP
Qwen3.8 27B: The new KING of local CODE models
DeepSeek V4 Pro: 50x cheaper than FABLE 5
Did GLM-5.3 reach Kimi K3 level with just post-training?
Muse Glimmer: Meta voltou ao Open Source com um 30B que roda LOCAL
Opus 5 vs GPT-5.6 Sol vs Kimi K3 vs GLM-5.2 vs MiniMax M3. Quem VENCE?
Is Muse Spark 1.2 worth your time?
DeepSeek V4 Flash LOCAL: Is it time to CANCEL Claude and GPT?
I built a REAL app with Qwen-3.6 27B LOCALLY without spending ANYTHING on tokens
Claude Opus 5 is the BEST model in the WORLD (beats even Fable 5)
Intel Arc Pro B50: A "Budget" RTX 2000 Pro with Only 70W
DeepSeek V4 Flash is BACK to take on OPUS
Qwen 3.8 Max Final Version is the BEST model of the YEAR
Qwen 3.8 Max claims it only loses to Fable 5, could it be?
AirLLM and Colibrì: Is this the FUTURE of LOCAL models?
Qwen 3.8 Max in agent mode: Opus-level results?
MiniMax H3: This Open Source Video AI Surprised Me
Qwen 3.7 Flash: Is it the BEST value-for-money model out there? (good, fast, and CHEAP)
Laguna-S-2.1 vs Ling-3.0-Flash: USA vs China Open Source. Which is BETTER?
Kimi K3, the Open Source model that GOES TOE-TO-TOE with GPT-5.6 and Fable 5
Is Bonsai 27B 1-bit the FUTURE of Local LLMs? (It even runs on PHONES)
Kimi K3 in agentic mode is IMPRESSIVE (Kimi CLI)
Is Grok 4.5 a CHEAP Fable?
GPT 5.6 SOL promised to be BETTER than Fable, and it DELIVERED (FABULOUS results)
GLM 5.2 on OpenCode: Is it the BEST Open Source model available today?
Fable 5 HUMILIATED Opus (SCARY results)
FABLE 5 is back, but with a MUCH bigger problem
Rio 3.5: The BRAZILIAN model that went viral worldwide and became a SCANDAL in 24h
Claude Fable 5: The BEST model in the WORLD has been released by Anthropic (Mythos quality level)
