cheapoS THE CHEAPSKATE CLUB
COMPUTE OBSERVATORY · TELEMETRY MACRO-DECK

The cheapoS
Engine Room.

Aggregate intelligence across installations. See which models builders deploy, how autonomous agents divide cognitive labor, and where zero-dollar compute replaces commercial retail API spend.

TOTAL COMMUNITY COMPUTE144,561,059Verified machine sync receipts
MODELS BENCHMARKED48across major compute gateways
AUTONOMOUS PERSONAS4Worker, Reviewer, Planner, Coordinator
COMMERCIAL BILLS AVOIDED~$273.81Maximum leverage across providers
COGNITIVE DIVISION OF LABOR

Autonomous Workshop Roles

How autonomous cheapoS agent swarms divide their cognitive workloads when executing software tasks.

🧑‍💻89.4% share

Worker

The execution workhorse: editing files, running terminal checks, and tool calling.

121.08M121,084,769 tokens
🛡️6.5% share

Reviewer

The quality gatekeeper: auditing diffs, inspecting tests, and enforcing safety.

8.81M8,813,400 tokens
🗺️4.1% share

Planner

The architect: breaking complex requests into actionable autonomous milestones.

5.51M5,508,886 tokens
🧭0.1% share

Coordinator

The dispatcher: routing tasks between subagents and orchestrating checkpoints.

75.2K75,183 tokens
Cognitive Ratio Benchmark: For every 10 tokens a Worker generates executing code, Reviewers audit ~13.7 tokens, and Planners structure ~0.7 tokens. Autonomous quality assurance accounts for ~6.5% of overall compute.
THE MODEL POWER RANKINGS

Community Model Index

Which models do software builders actually deploy when maximizing compute efficiency? Live telemetry aggregated across installations.

#Model NameProviderCompute ShareTotal TokensBills Eliminated
1
dots-studio/dots-3-note-previewopenrouter/dots-studio/dots-3-note-preview:free
OpenRouter
25.8%
37.35M~$112.04
2
gemini-3.1-flash-lite
Google
22%
31.81M~$9.54
3
openai/gpt-oss-120b
Other
11.3%
16.31M~$81.57
4
groq/qwen/qwen3.6-27b
Groq
10.2%
14.73M~$14.73
5
antigravity/gemini-3.1-flash-lite
Google
8%
11.58M~$3.47
6
dots-studio/dots-3-note-previewdots-studio/dots-3-note-preview:free
OpenRouter
4.2%
6.01M~$18.04
7
antigravity/gemini-3.7-flash-medium
Google
3.2%
4.57M~$1.37
8
qwen/qwen3.8-27b
Other
2.4%
3.52M~$3.52
9
cohere/north-mini-codeopenrouter/cohere/north-mini-code:free
OpenRouter
1.9%
2.78M~$0.84
10
qwen/qwen3.6-27b
Other
1.5%
2.23M~$2.23
11
cohere/north-mini-codecohere/north-mini-code:free
OpenRouter
1.5%
2.20M~$0.66
12
groq/qwen/qwen3.8-27b
Groq
1.2%
1.73M~$1.73
13
antigravity/gemini-3.7-flash-low
Google
1.1%
1.65M~$0.50
14
antigravity/gemini-3.1-pro-low
Google
1.1%
1.64M~$0.49
15
nvidia/google/gemma-4-31b-it
Google
0.8%
1.18M~$3.55
16
openai/gpt-oss-20b
Other
0.8%
1.11M~$3.33
17
groq/openai/gpt-oss-120b
Groq
0.7%
1.01M~$5.07
18
nvidia/nemotron-3-super-120b-a12bnvidia/nemotron-3-super-120b-a12b:free
OpenRouter
0.6%
926.3K~$4.63
19
nvidia/nemotron-3-super-120b-a12b
Other
0.3%
447.9K~$2.24
20
gemini-3.7-flash-low
Google
0.2%
300.2K~$0.09
21
ling-3.0-flash-fin-free
Other
0.2%
276.2K~$0.28
22
nvidia/nemotron-3-super-120b-a12bopenrouter/nvidia/nemotron-3-super-120b-a12b:free
OpenRouter
0.2%
263.3K~$1.32
23
gemma4:31b
Local
0.2%
262.0K~$0.79
24
google/gemma-4-31b-it
Google
0.1%
209.2K~$0.63
25
groq/openai/gpt-oss-20b
Groq
0.1%
151.1K~$0.45
26
inclusionai/ling-3.0-flash-fininclusionai/ling-3.0-flash-fin:free
OpenRouter
0.1%
101.4K~$0.10
27
poolside/laguna-xs-2.1poolside/laguna-xs-2.1:free
OpenRouter
0%
61.1K~$0.18
28
nvidia/nvidia/nemotron-3-super-120b-a12b
Other
0%
38.9K~$0.19
29
poolside/laguna-xs-2.1openrouter/poolside/laguna-xs-2.1:free
OpenRouter
0%
30.5K~$0.09
30
qwen2.5-coder:7b
Other
0%
25.4K~$0.08
31
auto/codingauto/coding:free
OpenRouter
0%
19.6K~$0.06
32
antigravity/gemini-3.7-flash-high
Google
0%
8.9K~$0.00
33
google/gemma-4-26b-a4b-itopenrouter/google/gemma-4-26b-a4b-it:free
Google
0%
888~$0.00
34
auto/best-free
Other
0%
879~$0.00
35
mistralai/mistral-nemotron
Other
0%
698~$0.00
36
google/gemma-4-26b-a4b-itgoogle/gemma-4-26b-a4b-it:free
Google
0%
555~$0.00
37
google/gemma-4-31b-itopenrouter/google/gemma-4-31b-it:free
Google
0%
555~$0.00
38
nvidia/nemotron-3-nano-omni-30b-a3b-reasoningnvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free
OpenRouter
0%
552~$0.00
39
nvidia/nemotron-3.5-lightningnvidia/nemotron-3.5-lightning:free
OpenRouter
0%
406~$0.00
40
nvidia/nemotron-3-ultra-550b-a55bnvidia/nemotron-3-ultra-550b-a55b:free
OpenRouter
0%
393~$0.00
41
nex-agi/nex-n2.5-pronex-agi/nex-n2.5-pro:free
OpenRouter
0%
362~$0.00
42
nex-agi/nex-n2.5-mininex-agi/nex-n2.5-mini:free
OpenRouter
0%
352~$0.00
43
inclusionai/ling-3.0-flash-santeinclusionai/ling-3.0-flash-sante:free
OpenRouter
0%
330~$0.00
44
inclusionai/ling-3.0-flash-vlinclusionai/ling-3.0-flash-vl:free
OpenRouter
0%
321~$0.00
45
google/gemma-4-31b-itgoogle/gemma-4-31b-it:free
Google
0%
222~$0.00
46
liquid/lfm-2.5-2.6bliquid/lfm-2.5-2.6b:free
OpenRouter
0%
203~$0.00
47
poolside/laguna-s-2.1openrouter/poolside/laguna-s-2.1:free
OpenRouter
0%
200~$0.00
48
openai/gpt-oss-safeguard-20b
Other
0%
195~$0.00
INFRASTRUCTURE PROVENANCE

Where Community Compute Originates

4 compute tiers working together: public free gateways, included developer quotas, local hardware, and pay-as-you-go frontier keys.

98.7%

Public Free Gateways

Free developer endpoints from Google AI Studio, Groq Cloud, OpenRouter free tiers, and Cerebras.

142,671,931 tokens
1.1%

Included Account Quotas

Pre-bundled developer environment quotas and zero-dollar editor allowances.

1,601,645 tokens
0.2%

Local Hardware

On-device inference powered by Apple Silicon unified memory (MLX), local Ollama, and CPU execution.

287,483 tokens
0.0%

Paid Pay-As-You-Go

Targeted out-of-pocket tokens reserved for high-stakes frontier reasoning or heavy benchmarks.

0 tokens

Want your model featured here?

Sponsor next month's spotlight engine on the observatory.

Sponsor prospectus ↗