AIpediya

Global AI models & tools catalog. Compare AI models, pricing, context, benchmarks and use cases. Find the right AI for any task.

[AIpediya @AIpediya]

Welcome to AIpediya! 🌐

Discover AI models and tools in one place.

Compare capabilities, pricing, context windows, independent benchmarks, and more.

Find the right AI for your next task.

🔗 https://t.co/z81FXxomoS

[AIpediya @AIpediya]

⏳ Gemini 4 Argon's intro API price has an end date: $2 / $10 per 1M input / output tokens through Nov 30. From Dec 1 it doubles to $4 / $20.

1,048,576-token context; takes text, image, video, audio and PDF.
https://t.co/FQfW14qSht

#Gemini #AIpricing

[AIpediya @AIpediya]

💸 The 5 cheapest LLM APIs on our price table right now, per 1M input / output tokens:

GPT-6 Luna: $0.10 / $0.50
GPT-4o Mini: $0.15 / $0.60
GPT-5.6 Luna: $0.20 / $1.20

Full top 5 in the card, all list prices here:
https://t.co/3EYMf5VkBG

#LLM #AIpricing

[AIpediya @AIpediya]

🎬 AI video with sound for $0.0425 per second?

FLUX 3 Max (fal Research + Black Forest Labs): text/image-to-video, 5–20s clips, 720p, optional 1080p upscale and generated audio. Promo price runs to Nov 7, then $0.17/s at 720p.

https://t.co/n7HlJEthww

#AIvideo #FLUX

[Dmytro Cherevko | AI & Dev @Cherevkodmit]

Пять голов лучше одной. Пять копий открытой Qwen3.8-27B через общий план и файлы на диске набрали 92,4% на LiveCodeBench против 90,4% у Claude Fable 5 (в одиночку — 66,8%). По сути ничья, но архитектура вместо масштаба — сильная идея. https://t.co/W244aGrIy2

[AIpediya @AIpediya]

@MiniMax_AI @tenstorrent Good to see more hardware options for M3.1 Flash. For reference, MiniMax's own API lists it at $0.15 / 1M input, $0.60 / 1M output and $0.015 / 1M cache-read tokens (checked Oct 21), a handy baseline when comparing hosted providers.

[AIpediya @AIpediya]

@togethercompute @reflection_ai @FactoryAI Useful detail for anyone evaluating Beam beyond the API beta: it's a sparse MoE with 501B total / 23B active params, an effective 1M-token context, and open weights under Apache 2.0 (BF16, FP8, NVFP4). Hosted per-token pricing isn't published yet.

https://t.co/IwpPYhfMKo

[AIpediya @AIpediya]

🧑‍💻 New from MiniMax: M3.1 Flash, a multimodal model for coding, agents and computer use.

API: $0.15 / $0.60 per 1M input / output tokens, cache reads $0.015. Weights are announced for a later release.

https://t.co/xnuCCvcpNc

#MiniMax #AIagents

[AIpediya @AIpediya]

⚡ An LLM that writes by diffusion, not token by token: Inception's Mercury 3.

Tool calling, structured output, 4 reasoning levels. Launch price (50% off, no end date published): $0.10 / $0.35 per 1M input / output tokens.

https://t.co/3LbPt4bYs0

#LLM #AI

[AIpediya @AIpediya]

🎧 Mistral's Voxtral 3 is out (Oct 20, API preview):

• Mini Transcribe 3: $0.0035/min
• Small Transcribe 3: $0.004/min
• Mini Transcribe Realtime 3: $0.008/min
• Mini TTS 3: $13 / 1M characters

Compare the cards:
https://t.co/G7yrMadaWX

#SpeechAI #MistralAI

[AIpediya @AIpediya]

🤖 FLUX for robots? Black Forest Labs' FLUX 3 Action: 7B open-weight policies for DROID and SO-101 robots. Camera + robot state + instruction in, actions + future frames out.

DROID success: 42.92% (developer-reported).
https://t.co/CtS3JW7PHe

#Robotics #OpenWeights

[AIpediya @AIpediya]

🔎 Building RAG? Cohere Embed 5 Fast vs Pro:

• Fast: $0.08 / 1M input tokens, for low-latency search
• Pro: $0.12 / 1M, for quality-critical indexing

Both take text + images, 128K context. Compare cards:
https://t.co/3G8DjypkiZ

#RAG #Embeddings

[AIpediya @AIpediya]

🤔 Haiku, Sonnet or Opus? Claude 5.5 API prices per 1M input/output tokens (standard):

• Haiku 5.5: $0.10 / $0.50 (prompts ≤100K)
• Sonnet 5.5: $2 / $10
• Opus 5.5: $4 / $20

All three: 1M-token context. Which runs your stack?
https://t.co/NbGJJNdltj

#Claude #LLM

[AIpediya @AIpediya]

🎙️ Describe a voice in words, get a custom AI voice.

OpenAI's GPT-Live 1 now has Prompt-to-Voice (/v1/audio/voices). Creating a voice is free; speech is billed at GPT-Live rates ($0.05/min audio layer). Plus 9 new voices.

https://t.co/w54FbDSTp2

#VoiceAI #OpenAI

[AIpediya @AIpediya]

📚 Drowning in papers and PDFs? Two research assistants to try:

🔹 NotebookLM (Google): answers & summaries grounded in the sources you upload
🔹 Elicit (Ought): web-based research assistant

https://t.co/M0Ldq3EY49
https://t.co/7KY45E9D34

Which do you use?
#AItools #Research

[AIpediya @AIpediya]

🧪 GPT-6.1 Sol, judged by independent tests, not marketing.

Its card lists 46 LiveBench results (e.g. 98% on AMPS_Hard math at max config), a 1.05M-token context and $2 / $10 per 1M input / output tokens (standard).

https://t.co/zRd8ByNE7n

#OpenAI #LLM

[AIpediya @AIpediya]

🌍 Compare AI models in your own language: AIpediya is available in 22 languages, from English and Español to 日本語, हिन्दी and Українська.

Cards are written in English & Russian and machine-translated into the rest.

Pick yours:
https://t.co/z81FXxomoS

#AI #AItools

[AIpediya @AIpediya]

🐋 DeepSeek V4.1 Flash got a new revision: V4.1-Flash-1017 (Oct 17), with weights published under MIT.

The API alias and pricing are unchanged: from $0.15 / $0.60 per 1M input / output tokens off-peak, with a 1M-token context.

https://t.co/xx89YkgDUG

#DeepSeek #OpenWeights

[AIpediya @AIpediya]

📈 Budget alert for Gemini 3.8 Flash users.

Standard API price: $0.75 / $3.75 per 1M input / output tokens through Dec 31, 2026. Google's announced rates from Jan 1, 2027: $1.50 / $7.50.

Plan your 2027 costs:
https://t.co/bP1UGvGIUk

#GeminiAPI #LLM

[AIpediya @AIpediya]

🔓 Reflection AI's Beam now has open weights.

Sparse MoE, 501B total / 23B active params, for coding, reasoning & agents. 1M-token effective context; Apache 2.0, in BF16, FP8 & NVFP4.

API is in public beta, price not published yet.
https://t.co/IwpPYhfeUQ

#OpenWeights #LLM

[AIpediya @AIpediya]

🪶 9B params, 2.00 GB of language weights.

PrismML's Ternary Bonsai 2 9B (released Oct 15): a Qwen3.5-based vision-language model with ternary weights (−1/0/+1), 262K context, Apache-2.0.

Note: needs PrismML's llama.cpp fork.
https://t.co/lK9aDmJXlU

#LocalAI #OpenWeights