Measured work. Comparable results.
Leaderboard
Find the configuration for your needs. Browse every tested configuration. Use comparison filters when you want the same tests and execution rules. How scoring works
All tested configurations. Latest run per model, execution method, effort and recorded hardware. Scores describe the pages produced; different prompts and workloads are not a controlled model ranking. Browse every run
Order by scoreModel / configurationScore / basisCoverageResults
1
gpt-6-astraOpenAI · Codex CLI (agent) · high
7.5Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- gpt-6-astra
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 3 Oct 2026
2
deepseek/deepseek-v4.1-flash-1017DeepSeek · OpenRouter (Pi agent) · high
7.5Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- deepseek/deepseek-v4.1-flash-1017
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 2 · blast-island: 2 · camera-shop: 2 · jet-canyon: 2 · launch-day: 2
- Tested
- 17 Oct 2026
3
claude-opus-5-5Anthropic · Claude Code (agent) · high
7.4Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- claude-opus-5-5
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 3 Oct 2026
4
step-5-preview-freeStepFun · OpenCode Go (Pi agent) · high
7.4Quality v1 · 2/2 pages2/2scored pagesConfiguration details
- Model ID
- step-5-preview-free
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 0 · blast-island: 0 · camera-shop: 1 · jet-canyon: 0 · launch-day: 0
- Tested
- 8 Oct 2026
5
claude-sonnet-5-5Anthropic · Claude Code · default effort
7.3Quality v1 · 4/5 pages4/5scored pagesConfiguration details
- Model ID
- claude-sonnet-5-5
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 2
- Tested
- 29 Sept 2026
6
claude-sonnet-5-5Anthropic · Claude Code (agent) · high
7.3Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- claude-sonnet-5-5
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 2 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 2
- Tested
- 5 Oct 2026
7
gemini-4-argonGoogle · Gemini API (Pi agent) · high
7.2Quality v1 · 4/5 pages4/5scored pagesConfiguration details
- Model ID
- gemini-4-argon
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 13 Oct 2026
8
meta/muse-wonderMeta · OpenRouter (Pi agent) · max
6.9Quality v1 · 5/6 pages5/6scored pagesConfiguration details
- Model ID
- meta/muse-wonder
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 12 Oct 2026
9
stepfun/step-5-previewStepFun · OpenRouter (Pi agent) · high
6.5Quality v1 · 6/8 pages6/8scored pagesConfiguration details
- Model ID
- stepfun/step-5-preview
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 8 Oct 2026
10
qwen3.8-flash-next-gsq-rcoQwen · LM Studio (Pi agent) · default effort
6.3Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- qwen3.8-flash-next-gsq-rco
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 5 Oct 2026
11
gpt-6.1-solOpenAI · Codex CLI · default effort
6.3Quality v1 · 5/5 pages5/5scored pagesConfiguration details
- Model ID
- gpt-6.1-sol
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- Historical policy unavailable
- Tested
- 29 Sept 2026
12
gpt-6-lunaOpenAI · Codex CLI (agent) · max
6.2Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- gpt-6-luna
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 2 Oct 2026
13
claude-haiku-5-5Anthropic · Claude Code (agent) · high
6.1Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- claude-haiku-5-5
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 4 · blast-island: 3 · camera-shop: 1 · jet-canyon: 3 · launch-day: 2
- Tested
- 7 Oct 2026
14
deepseek-v4.1-flash:cloudDeepSeek · Ollama · default effort
5.9Quality v1 · 5/5 pages5/5scored pagesConfiguration details
- Model ID
- deepseek-v4.1-flash:cloud
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 2 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 30 Sept 2026
15
glm-5.3-flash:cloudZ.ai · Ollama (Pi agent) · high
5.4Quality v1 · 7/8 pages7/8scored pagesConfiguration details
- Model ID
- glm-5.3-flash:cloud
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 4 Oct 2026
16
kimi-k3:cloudMoonshotAI · Ollama · default effort
4.5Quality v1 · 8/8 pages8/8scored pagesConfiguration details
- Model ID
- kimi-k3:cloud
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 2 · launch-day: 1
- Tested
- 3 Oct 2026
17
openai/gpt-6-lunaOpenAI · OpenRouter API · default effort
3.6Quality v1 · 5/5 pages5/5scored pagesConfiguration details
- Model ID
- openai/gpt-6-luna
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- Historical policy unavailable
- Tested
- 27 Sept 2026
18
gpt-6-lunaOpenAI · Codex CLI (agent) · high
3.5Quality v1 · 1/1 pages1/1scored pagesConfiguration details
- Model ID
- gpt-6-luna
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 0 · camera-shop: 0 · jet-canyon: 0 · launch-day: 0
- Tested
- 2 Oct 2026
19
x-ai/grok-4.7SpaceXAI · OpenRouter (agent) · default effort
3.4Quality v1 · 1/2 pages1/2scored pagesConfiguration details
- Model ID
- x-ai/grok-4.7
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 0 · camera-shop: 0 · jet-canyon: 1 · launch-day: 0
- Tested
- 3 Oct 2026
20
unbiased/pareto-26.10-previewUnbiased · OpenRouter API · default effort
2.2Quality v1 · 5/5 pages5/5scored pagesConfiguration details
- Model ID
- unbiased/pareto-26.10-preview
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 1 Oct 2026
21
qwen3.8:27b-mlxQwen · Ollama · default effort
2.1Quality v1 · 5/5 pages5/5scored pagesConfiguration details
- Model ID
- qwen3.8:27b-mlx
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 1 · launch-day: 1
- Tested
- 1 Oct 2026
22
Kolibri-1-MLX-4bitOther · MLX · default effort
1.8Quality v1 · 7/8 pages7/8scored pagesConfiguration details
- Model ID
- ~/.lmstudio/models/velaia/Kolibri-1-MLX-4bit
- Size
- Undisclosed
- Quantisation
- Undisclosed
- Hardware
- Undisclosed
- Attempts by test
- aero-lab: 1 · blast-island: 1 · camera-shop: 1 · jet-canyon: 2 · launch-day: 1
- Tested
- 4 Oct 2026