Shared benchmarks

31
Shared benchmarks for Gemini 3.5 Flash and Gemini 3.6 Flash
BenchmarkGemini 3.5 FlashGemini 3.6 Flash
  • 27.5%
  • 46.9%
Evaluation details for APEX-Agents Original

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on APEX-Agents Original

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on APEX-Agents Original
  • 9.0%low
  • 11.4%medium
  • 13.4%high
  • 15.7%low
  • 16.4%medium
  • 17.1%high
Evaluation details for AutomationBench 1.0.6

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on AutomationBench 1.0.6
Record for Gemini 3.5 Flash on AutomationBench 1.0.6
Record for Gemini 3.5 Flash on AutomationBench 1.0.6

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on AutomationBench 1.0.6
Record for Gemini 3.6 Flash on AutomationBench 1.0.6
Record for Gemini 3.6 Flash on AutomationBench 1.0.6
  • 43.3%
  • 54.6%
Evaluation details for BrowseComp 130-question subset — Mercor web research agent

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 94.3%
  • 95.1%
Evaluation details for CharXiv Descriptive and reasoning — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 84.2%
  • 85.2%
Evaluation details for CharXiv Reasoning — No tools

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on CharXiv Reasoning — No tools

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on CharXiv Reasoning — No tools
  • 84.9%
  • 89.4%
Evaluation details for CharXiv Reasoning — Search and code execution

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on CharXiv Reasoning — Search and code execution

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on CharXiv Reasoning — Search and code execution
  • 36.0%
  • 47.0%
Evaluation details for DeepSWE 1.1

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on DeepSWE 1.1

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on DeepSWE 1.1
  • 36.6%
  • 49.3%
Evaluation details for DeepSWE 1.1 — Mercor mini-swe-agent, 500 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 57.9%
  • 56.3%
Evaluation details for Finance Agent Benchmark 2 — October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 88.9%low
  • 92.8%high
  • 86.4%low
  • 94.1%high
Evaluation details for GPQA Diamond — Epoch AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 85.7%
  • 93.4%
Evaluation details for GPQA Diamond — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 92.7%
  • 93.4%
Evaluation details for GPQA Diamond — Vals AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness
  • 2.5%
  • 3.3%
Evaluation details for HLAB October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 40.6%
  • 41.2%
Evaluation details for Humanity's Last Exam Text only — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 83.5%
  • 82.8%
Evaluation details for MedXpertQA MM — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on MedXpertQA MM — Mercor single-shot

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on MedXpertQA MM — Mercor single-shot
  • 90.0%
  • 90.0%
Evaluation details for MMLU-Pro Original — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 89.5%
  • 89.3%
Evaluation details for MMLU-Pro Original — Vals AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 84.7%
  • 83.7%
Evaluation details for MMMU-Pro 1,730 questions — Mercor single-shot

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on MMMU-Pro 1,730 questions — Mercor single-shot

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on MMMU-Pro 1,730 questions — Mercor single-shot
  • 88.3%
  • 88.4%
Evaluation details for MMMU-Pro Vals AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on MMMU-Pro Vals AI harness

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on MMMU-Pro Vals AI harness
  • 78.4%
  • 83.0%
Evaluation details for OSWorld Verified

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on OSWorld Verified

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on OSWorld Verified
  • 36.8%
  • 30.4%
Evaluation details for Riemann-bench October 2026 leaderboard

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 66.2%
  • 66.2%
Evaluation details for SimpleQA Verified Original — Epoch AI harness

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 19.6%
  • 32.0%
Evaluation details for SWE Atlas Codebase QnA — Mercor mini-swe-agent, 250 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 68.2%
  • 75.1%
Evaluation details for SWE-bench Multilingual — Mercor mini-swe-agent, 1000 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

  • 78.5%
  • 81.1%
Evaluation details for SWE-bench Verified — Mercor mini-swe-agent, 1000 steps

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 78.8%
  • 79.6%
Evaluation details for SWE-bench Verified — Vals AI mini-swe-agent

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Gemini 3.6 Flash

  • 55.1%
  • 58.7%
Evaluation details for SWE-bench Pro Public

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on SWE-bench Pro Public

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on SWE-bench Pro Public
  • 76.2%
  • 78.0%
Evaluation details for Terminal-Bench 2.1

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on Terminal-Bench 2.1

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on Terminal-Bench 2.1
  • 73.8%
  • 74.5%
Evaluation details for Terminal-Bench 2.1 — Mercor Terminus 2

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on Terminal-Bench 2.1 — Mercor Terminus 2

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on Terminal-Bench 2.1 — Mercor Terminus 2
  • 74.2%
  • 73.8%
Evaluation details for Terminal-Bench 2.1 — Vals AI Terminus 2

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.

Gemini 3.5 Flash

Record for Gemini 3.5 Flash on Terminal-Bench 2.1 — Vals AI Terminus 2

Gemini 3.6 Flash

Record for Gemini 3.6 Flash on Terminal-Bench 2.1 — Vals AI Terminus 2
  • 48.7%
  • 64.0%
Evaluation details for Vibe Code Bench 1.1 — OpenHands, October 2026 evaluation

The recorded benchmark version, metric, and evaluator sets match. Evaluation methodology is not recorded by the registry.