← GPT-5.6 Sol · API · Max effort
Evidenceacceptedthird_party_leaderboard
Chatbot Arena / LMArena human preference
Chatbot Arena leaderboard mirror (openlm.ai) — GPT-5.6 Sol
ev_2026_07_gpt56sol_arena
Result
1509
Confidence B
- Tested
- 2026-07-11
- Runs
- 1
Configuration
How it was measured
- Exact model
- gpt-5.6-sol
- Reasoning
- max
- Agent harness
- none
- Tool access
- chat product surface (not tool-agent harness)
- Prompting
- anonymous pairwise battles (Arena)
PRIMARY CHA — blind human preference Elo on Chatbot Arena style leaderboard (openlm.ai mirror snapshot). Sol is ~2 days post-GA; vote counts may still be thinner than mature peers.
Caveats
Known limitations
- Preference Elo ≠ precise constraint following (pair with IF when available).
- Post-GA vote count may still be thin; Elo can move as sample grows.
- Mirror site; LMArena primary shows slightly lower xhigh row (~1486).
- Product chat surface may differ from API max-effort coding configs.
Weighting
Evidence weight factors
The engine weights each evidence row by these factors (0–1) when fusing scores — higher is more trusted.
relevance0.90
quality0.85
recency1.00
health0.85
comparability0.75
reliability0.75
Benchmark
Chatbot Arena / LMArena human preference
Blind pairwise human preference battles aggregated into Elo ratings for chat quality, style, and helpfulness under real user prompts.
Benchmark source →