AI looks like software. Underneath, it is a system of chips, power, models, labor and capital. Choose a thread and follow it all the way down.
562 community Jev demos, labeled by Jev: classifiers and agents became real products, games only earned views. 39 examples with reported cost and speed.
Read the analysisPhoto: HackTX 2012 More Participants by Randymcking, CC BY-SA 3.0, via Wikimedia Commons. Cropped.
Four readings, each dated and linked to the tracker it came from.
18shipped in October 2026
13 labs. Newest: Muse Wonder from Meta.
MicrosoftCloudflare+7 moreAs of Oct 13, 2026
$456 to $590$14.25 to $18.44 per GB
Plus DDR4, HBM and server RDIMM street prices.
As of Oct 7, 2026
50models with verified rates
1 more announced or rumored. Every rate is source-linked.
As of Oct 13, 2026
117benchmarks, 12 categories
86 still active, the rest saturated or retired.
As of Oct 1, 2026
Claude Haiku 5.5 costs $0.10/$0.50 per million tokens for prompts under 100K, 90% below Haiku 4.5, and scores 43 on Artificial Analysis at $0.21 a task.
Not for chat. For agents that outlive a 300-second function, yes, or AI SDK 7 WorkflowAgent instead. Where TanStack AI fits, and what durability costs per run.
Mistral Large 4 costs $0.68/$2.09 per million tokens on launch sale, $1.36/$4.18 at list. Artificial Analysis scores it 38. Weights are due by end of October.
Reflection AI's Beam is a 501B open-weight model, 23B active, with Apache 2.0 weights due in October 2026. Its own benchmarks put it level with GLM-5.2.
OpenUI is the MIT-licensed generative UI framework from Thesys. Its OpenUI Lang streams 2.1x fewer tokens than JSON, not 3x. The hype, use cases, and rivals.
Eight of 13 AI agent platforms checked in October 2026 support BYOK: Clay, Gumloop, n8n, Sim and more. Your own key rarely cuts the bill, and here is why.
Check whether Claude got worse: independent evidence for each Claude model, a live daily benchmark chart, and how a nerf test works and what it misses.
Frontier LLMs score 95 to 98% of their English level in about a dozen major languages. Below that, quality falls fast. A four-tier map with the evidence.
Three pictures from the same datasets the trackers run on. Each one is drawn at build time from a dated snapshot, not a live feed.
| Model | Intelligence Index | Blended price per Mtok | On the cost-efficiency frontier |
|---|---|---|---|
| Claude Opus 5.5 | 58 | $8.00 | Yes |
| Claude Sonnet 5.5 | 56 | $4.00 | Yes |
| GPT-6 Astra | 53 | $20.00 | No |
| Gemini 4 Argon | 53 | $4.00 | No |
| GPT-6.1 Sol | 52 | $4.00 | No |
| Muse Wonder | 52 | $8.00 | No |
| Muse Spark 1.3 | 48 | $2.00 | Yes |
| Grok 4.7 | 46 | $3.00 | No |
| GLM-5.3 | 45 | $2.20 | No |
| Qwen3.8 Max | 45 | $3.00 | No |
| Kimi K3 | 44 | $6.00 | No |
| Claude Haiku 5.5 | 43 | $0.20 | Yes |
| GLM-5.3-Flash | 42 | $0.24 | No |
| DeepSeek V4.1 Flash | 39 | $0.52 | No |
| Mistral Large 4 | 38 | $1.03 | No |
| MiniMax M3 | 29 | $0.50 | No |
| Inkling | 25 | $2.57 | No |
| Nemotron 3 Ultra | 23 | $1.40 | No |
| Date | Typical listing | Cheapest tracked |
|---|---|---|
| 2025-04-04 | $125 | $59 |
| 2025-06-11 | $129 | $55 |
| 2025-08-01 | $132 | $54 |
| 2025-09-21 | $132 | $63 |
| 2025-11-11 | $274 | $98 |
| 2026-01-01 | $483 | $195 |
| 2026-02-21 | $579 | $195 |
| 2026-04-13 | $546 | $390 |
| 2026-06-03 | $576 | $201 |
| 2026-07-24 | $573 | $390 |
| 2026-09-06 | $586 | $420 |
| 2026-09-09 | $585 | $420 |
| 2026-09-12 | $577 | $425 |
| 2026-09-15 | $581 | $425 |
| 2026-09-18 | $584 | $429 |
| 2026-09-21 | $602 | $429 |
| 2026-09-24 | $603 | $429 |
| 2026-10-07 | $592 | $456 |
| Date | Cheapest GPT-4-class model | Frontier flagship |
|---|---|---|
| 2021-11-01 | $60 | |
| 2023-03-01 | $60 | $60 |
| 2024-05-01 | $15 | |
| 2024-07-01 | $0.60 | |
| 2024-12-01 | $60 | |
| 2026-07-01 | $0.40 | $30 |
| 2026-08-21 | $20 | |
| 2026-09-03 | $50 |
Every number checked, dated, and linked to its source.
The price you are shown is never the price you pay. Your flat monthly plan is still metering tokens behind the scenes. Your coding agent burns most of its budget on tasks it never finishes. The data center sold in gigawatts arrives three years late, stuck in line for electricity.
This site works out the real figure and shows you where it came from.
Six subjects, one question: what does this actually cost to run?
The free trackers handle the moving parts. Model prices, GPU and memory street prices, benchmark definitions, provider rates, cost per task: each one is a dated dataset, read again at the source whenever the market shifts. You will never be handed a screenshot from six months ago.
Nothing on this site is guessed. A benchmark score comes off the leaderboard, not out of a launch post. A price links to the page it was read from. No ranking is for sale, and any paid link is labelled as one. Anything that turns out to be wrong gets corrected in the open.
Every figure links to its source right where it appears, not in a footnote at the bottom. The data behind the trackers is published as JSON, so you can cite it directly instead of scraping the page. Each post says how many sources it stands on and when those numbers were last checked.
The label matters as much as the link. A preprint is called a preprint. Documentation written by a vendor is called documentation, not independent proof. Coverage from another outlet says whose reporting it is.
Prices and scores are stamped with the date they were confirmed. A rate checked in March proves nothing in August, and leaderboards reshuffle faster than any page can keep up. When a number is estimated rather than measured, the assumptions are laid out so you can redo the math, or argue with it.
