6.1 Sol ultrafast >>> 6 Astra ~ same cost but whole different game with 250 tpm
ECONOMY
-
SubQ Snap Pricing Enables Large Multimodal Document Decision Workflows
By
–
WAIT… $0.05 PER MILLION INPUT TOKENS?!
— Charly Wargnier ♨️ (@DataChaz) 12 octobre 2026
And output is free.
SubQ just launched Snap with up to 384K tokens of multimodal context and reported sub-second decisions on 100K tokens.
That pricing makes experimenting with document-heavy workflows seriously tempting 🔥
+ Free 100M… https://t.co/ggwRz23qlD pic.twitter.com/3CRTuXq2iVWAIT… $0.05 PER MILLION INPUT TOKENS?! And output is free. SubQ just launched Snap with up to 384K tokens of multimodal context and reported sub-second decisions on 100K tokens. That pricing makes experimenting with document-heavy workflows seriously tempting 🔥 + Free 100M
-
Muse Wonder Pricing Compares Favorably With Fable’s Standard Token Rates
By
–
The crazy thing is that Meta Muse wonder outperforms Fable 5.1 at 60% cheaper pricing. Again:Claude Fable 5 and 5.1 cost $10 per million input tokens and $50 per million output tokens, making Muse Wonder’s reported $4/$20 pricing 60% cheaper for both.
-
Muse Wonder Listing Reveals Million-Token Context and Token Pricing
By
–
Here we go: Metas new model already spotted: Muse Wonder, Metas Fable Class contender. @synthwavedd already found out that it appears to be rolling out at $4 per million input tokens and $20 per million output tokens, with a 1M-token context window. “Wonder Full,” its
-
SubQ Snap Combines Large Multimodal Context With Subsecond Decisions
By
–
Subquadratic says SubQ Snap processes 100K tokens and returns a decision in *under a second*.
— Chubby♨️ (@kimmonismus) 12 octobre 2026
Its new decision model supports up to 384K tokens in one call, including text and images. The team’s pricing is $0.05 per million input tokens, cached or uncached, with free output.… https://t.co/ezAysGOnw7Subquadratic says SubQ Snap processes 100K tokens and returns a decision in *under a second*. Its new decision model supports up to 384K tokens in one call, including text and images. The team’s pricing is $0.05 per million input tokens, cached or uncached, with free output.
-
AI Token Generation Capacity Could Remain a Compute Bottleneck
By
–
We don’t have enough compute for all the things AI would allow us to do btw Tokens generation will be the bottleneck for years to come
-
Perplexity CEO Describes High-Spend Solopreneur Automation Business
By
–
FYI: Aravind Srinivas (@AravSrinivas) says there’s a 1-person company spending **~$1M/year** on Perplexity
— Molly O’Shea (@MollySOShea) 12 octobre 2026
“There’s 1 solopreneur, literally a one-person company, spending upwards of a million dollars a year on our product.”
“He’s making $3 million a year by automating… https://t.co/LLhTQTyWks pic.twitter.com/iii8R7pqMaFYI: Aravind Srinivas (
@AravSrinivas
) says there’s a 1-person company spending **~$1M/year** on Perplexity “There’s 1 solopreneur, literally a one-person company, spending upwards of a million dollars a year on our product.” “He’s making $3 million a year by automating -
OpenAI Decisions API Returns Choices, Probabilities and Scores
By
–
For builders: OpenAI's Decisions API is in public beta. It returns a choice, a probability or a score in place of paragraphs, handy for routing, tagging and quick yes/no calls. Runs on GPT-6 Luna, with input-only pricing at $0.10 per million tokens.
-
Haiku 5.5 Adds Adjustable Effort and Lower Inference Costs
By
–
Anthropic released Claude Haiku 5.5, the small, fast member of the 5.5 family. Anthropic says it costs around 75% less to run than Haiku 4.5, and it's the first Haiku with an adjustable effort setting. Sonnet 5.5 cache reads were also cut in half.https://t.co/fUnsAyKoQc…
— Futurepedia – Learn to Leverage AI (@futurepedia_io) 12 octobre 2026Anthropic released Claude Haiku 5.5, the small, fast member of the 5.5 family. Anthropic says it costs around 75% less to run than Haiku 4.5, and it's the first Haiku with an adjustable effort setting. Sonnet 5.5 cache reads were also cut in half. https://
anthropic.com/claude-haiku-5
-5
… -
Rising H100 Rental Prices Suggest Persistent AI Compute Shortages
By
–
Kind of nuts, H100 price went up from $1.8/h to $2.8/h despite Blackwells, Vera Rubins etc. which are far many times more efficient. Compute shortage is not easing at all.