GPQA: Google-Proof Q&A
PhD-level multiple-choice questions in biology, physics, and chemistry designed so that even experts with internet access score below 67%. The Diamond subset (198 questions) is the hardest variant used in most frontier model evaluations.
Models ranked
209
tracked on this benchmark
Score band
95.8 – 25.2
best → lowest tracked
Snapshot trend
+16.18
Oct 14 → Oct 16 · 1 models
Leaderboard
Tracked models ranked by Accuracy (higher is better).
Configuration: gpt-6-astra (Max)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'gpt-6-astra_max', run started 2026-08-30). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Configuration: claude-sonnet-5-5 (Max)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'claude-sonnet-5-5_max', run started 2026-09-29). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Configuration: gemini-3.8-flash (High)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'gemini-3.8-flash_high', run started 2026-09-02). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Configuration: gpt-6.1-sol (Max)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'gpt-6.1-sol_max', run started 2026-09-29). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Configuration: gemini-3.7-flash (High)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'gemini-3.7-flash_high', run started 2026-08-14). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Notes: Confidence: medium. weekly best-of refresh: multiple secondary trackers agree on 94.6%; pending an Anthropic primary score.
Notes: Official GPT-5.6 GA launch benchmark row. Uses the repo canonical gpqa slug for the GPQA Diamond variant.
Configuration: gpt-6-sol (Max)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'gpt-6-sol_max', run started 2026-09-22). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Configuration: gemini-3.6-flash (High)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'gemini-3.6-flash_high', run started 2026-08-02). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Configuration: grok-4.6 (High)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'grok-4.6_high', run started 2026-08-12). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Notes: Uses the repo canonical gpqa slug for the GPQA Diamond variant; do not create a duplicate gpqa-diamond benchmark.
Notes: Official Anthropic launch benchmark using the repo canonical gpqa slug for GPQA Diamond.
Notes: Model variant: Kimi K3 (max reasoning effort). Benchmark variant: GPQA-Diamond. Harness/evaluator not disclosed by Moonshot. Source posture: Moonshot/Kimi self-reported launch table; confidence: medium because the value is vendor-reported and not independently reproduced. Recommended seed value: 93.5.
Configuration: grok-4.7 (Extra High)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'grok-4.7_xhigh', run started 2026-09-22). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
Configuration: qwen3.8-max-0902 (Extra High)
Notes: Epoch AI-run evaluation (evaluator of record), from Epoch AI Benchmarking Hub data download https://epoch.ai/data/benchmark_data.zip (file gpqa_diamond.csv, row 'qwen3.8-max-0902_xhigh', run started 2026-09-02). Epoch data is CC BY 4.0: credit 'Epoch AI, Capabilities & benchmarking, https://epoch.ai/benchmarks'. Highest-scoring Epoch configuration for this model selected; other configurations exist in the same file. Discovered via Benchmark Heaven (lead only).
How to read this benchmark
This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.
Trust this score when
- There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
- The model list covers the same version family you can actually deploy today.
- Top candidates overlap with your required routing and feature requirements.
Be cautious when
- There is only one benchmark snapshot or the dataset appears stale.
- The benchmark metric direction is opposite of your decision objective.
- The score difference between options is narrow and likely within implementation variance.
Related benchmarks
Last reviewed: Apr 15, 2026