OpenGPTEvidence-based model decisionsStarting candidate 1
StableClaude Fable 5
claude-fable-5AnthropicGeneral capability · #1
03 · TASK RANKING
Deep research
Multi-step analysis, evidence synthesis, and reasoned conclusions.
- Official version
- 3
- Public evidence
- 3
- Data cutoff:
- Why these candidates
- Start with models covered by current intelligence and objective-task sources, then test citation discipline and uncertainty.
- What remains uncertain
- Benchmark coverage does not prove source quality, browsing quality, or factual reliability in your domain.
- Public evidence
- General capability
- Data cutoff
Starting candidate
3 / 3Three official versions to test first
Public data narrows the field. It does not replace a controlled test on your own work.
Starting candidate 2
StableGPT-5.6 Sol
gpt-5.6-solOpenAIGeneral capability · #2
Starting candidate 3
StableKimi K3
kimi-k3Moonshot AIGeneral capability · #4
Public evidenceThe models use different public evidence scales. Missing evidence is not treated as zero, and OpenGPT does not combine unlike sources into one score.⌄
Claude Fable 5
- Source
- Artificial Analysis ↗
- Source rank
- #1
- Source value
- 60
- Mapping reviewed
GPT-5.6 Sol
- Source
- Artificial Analysis ↗
- Source rank
- #2
- Source value
- 59
- Mapping reviewed
Kimi K3
- Source
- Artificial Analysis ↗
- Source rank
- #4
- Source value
- 57
- Mapping reviewed
A practical starting point, not a universal winnerSelect two or three candidates for a side-by-side public data comparison.
The models use different public evidence scales. Missing evidence is not treated as zero, and OpenGPT does not combine unlike sources into one score.
Benchmark coverage does not prove source quality, browsing quality, or factual reliability in your domain.
The evaluation opens with this task template and the first two selected models. You can edit every task before collecting answers.