All benchmarks

Internal Research Debugging Evaluation

Paper

Progress Over Time

Interactive timeline showing model performance evolution on Internal Research Debugging Evaluation

State-of-the-art frontier
Open
Proprietary

Internal Research Debugging Evaluation Leaderboard

3 models
ContextCostLicense
1
OpenAI
GPT-5.6 SolNew
OpenAI
—1.1M$5.00 / $30.00
2
OpenAI
GPT-5.6 TerraNew
OpenAI
—1.1M$2.50 / $15.00
3
OpenAI
GPT-5.6 LunaNew
OpenAI
—1.1M$1.00 / $6.00
Notice missing or incorrect data?
About this benchmark

What is Internal Research Debugging Evaluation?

The Internal Research Debugging Evaluation measures whether models can debug 41 real bugs from internal OpenAI research experiments (plus alignment-auditing tasks), where the original solutions took experienced researchers hours to days. Passing corresponds to providing assistance that would unblock the user, including partial root-cause explanations or fixes.

Internal Research Debugging Evaluation is a text benchmark evaluating models on reasoning, agents, and code tasks. LLM Stats tracks 3 models on this benchmark, scored on a 0–1 scale. The current average is 0.6, with the leader at 0.7.

Compare leaders on the best AI for reasoning, best AI for agents and best AI for code leaderboards.

Current leaders

GPT-5.6 Sol from OpenAI currently leads the Internal Research Debugging Evaluation leaderboard with a score of 0.683 across 3 evaluated AI models.

1GPT-5.6 SolOpenAI68.3%
2GPT-5.6 TerraOpenAI67.8%
3GPT-5.6 LunaOpenAI50.8%

FAQ

Common questions about the Internal Research Debugging Evaluation benchmark and leaderboard.

What is the Internal Research Debugging Evaluation benchmark?

The Internal Research Debugging Evaluation measures whether models can debug 41 real bugs from internal OpenAI research experiments (plus alignment-auditing tasks), where the original solutions took experienced researchers hours to days. Passing corresponds to providing assistance that would unblock the user, including partial root-cause explanations or fixes.

What is the Internal Research Debugging Evaluation leaderboard?

The Internal Research Debugging Evaluation leaderboard ranks 3 AI models based on their performance on this benchmark. Currently, GPT-5.6 Sol by OpenAI leads with a score of 0.683. The average score across all models is 0.623.

What is the highest Internal Research Debugging Evaluation score?

The highest Internal Research Debugging Evaluation score is 0.683, achieved by GPT-5.6 Sol from OpenAI.

How many models are evaluated on Internal Research Debugging Evaluation?

3 models have been evaluated on the Internal Research Debugging Evaluation benchmark, with 0 verified results and 3 self-reported results.

Where can I find the Internal Research Debugging Evaluation paper?

The Internal Research Debugging Evaluation paper is available at https://deploymentsafety.openai.com/gpt-5-6. The paper details the methodology, dataset construction, and evaluation criteria.

What categories does Internal Research Debugging Evaluation cover?

Internal Research Debugging Evaluation is categorized under reasoning, agents, and code. The benchmark evaluates text models.

Which model offers the best value on Internal Research Debugging Evaluation?

Among models scoring within 10% of the leader, GPT-5.6 Terra from OpenAI is the cheapest, at $2.50 per million input tokens with a score of 0.678.

How recent are the Internal Research Debugging Evaluation leaderboard results?

The Internal Research Debugging Evaluation leaderboard was last updated in July 2026 and currently includes 3 evaluated models.