CLARITY (BIRD) - Lexical Ambiguity (Multi-Facet): leaderboard

CLARITY (2026) builds ambiguous and unanswerable text-to-SQL requests on BIRD and resolves them through generated user-agent clarification dialogues; the model reads the dialogue and writes one SQL query, zero-shot at temperature 0. This board covers requests with two or more lexically ambiguous column terms. Execution accuracy in percent; higher is better.

Metric: Execution accuracy (%) on multi-turn BIRD instances whose request has two or more lexically ambiguous column terms; the model reads a generated user-agent clarification dialogue and writes one SQL query; execution accuracy against the original query, macro-averaged over conversation types; zero-shot, temperature 0, 2,048 output tokens; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 5 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-4.172.4#506
2GPT-4o70.8#754
3GPT-567.6#185
4Grok 3 Mini Fast66.2#681
5Llama 3.3 70B Instruct64.4#1394

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=clarity-bird-lexical-ambiguity-multi-facet · How It Works · Data refreshed daily, snapshot 2026-11-22.