Lava lamp v1 · OpenCode21.11.2026 07:24
Inkling 2 Mini
Low effortWorks
8 of 8 checks passed. All 5 required checks passed.
What it built
Build time
4m 07s
16th fastest of 18 models at low effort.
Output tokens
13,851
14th fewest of 18 models at low effort.
Output speed
56 tok/s
11th highest of 18 models at low effort.
API cost
None
The model ran on our own GPU: no API, so no cost.
Same model
Inkling 2 Mini at every effort.
Scroll sideways for all 4
Same effort
Other models at low effort.
Fastest first. This run took 4m 07s.
Scroll sideways for all 17
Muse Ember23.2 sWorks
10.7× faster than this run
GPT-5.6 Luna29.7 sWorks
8.3× faster than this run
GPT-6.1 Terra42.9 sWorks
5.8× faster than this run
GPT-6 Luna51.5 sFails
4.8× faster than this run
GPT-5.6 Terra1m 38sWorks
2.5× faster than this run
Claude Opus 5.51m 42sWorks
2.4× faster than this run
GPT-5.6 Sol1m 46sWorks
2.3× faster than this run
Claude Haiku 5.52m 36sWorks
1.6× faster than this run
Gemini 4 Neon2m 36sWorks
1.6× faster than this run
Claude Sonnet 5.52m 37sWorks
1.6× faster than this run
GPT-6 Sol2m 39sWorks
1.5× faster than this run
Claude Fable 5.12m 50sWorks
1.5× faster than this run
GPT-6.2 Sol2m 56sWorks
1.4× faster than this run
GPT-6.1 Sol3m 11sWorks
1.3× faster than this run
GPT-6.2 Astra3m 54sWorks
1.1× faster than this run
GPT-6 Astra4m 17sWorks
about the same
Claude Fable 5.24m 43sWorks
1.1× slower than this run
Checks
8 of 8 checks passed.
- Self-contained index.html is presentRequired
- Visible page contentRequired
- Visible autonomous animationRequired
- Accessible on/off controlOptional
- 390 px mobile layout has no horizontal overflowOptional
- Local animation frame performanceOptional
- No browser errorsRequired
- No attempted external or linked-resource requestsRequired
It works when it passes every required check; optional checks are reported, but they don’t decide that.
Run details
- Agent
- OpenCode
- Model
- Inkling 2 Mini
- Effort
- Low
- Build time
- 4m 07s
- Start to finish
- 5m 06s
- First output
- 2m 41s
- Input tokens
- 178,707 (162,325 from cache)
- Output tokens
- 13,851
- API cost
- None: ran on one NVIDIA GeForce RTX 3090 24 GB
Full record (run.json) Every check, hash, version and revision, as published.


