BlogLLM Benchmark Scores as Published, Side by Side
Gemini 3 Pro: Benchmark Scores as Published
Every result the catalog holds for Gemini 3 Pro: 110 tests in 14 areas, from 12 pages. 19 printed by Google itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer GoogleReleased 2025-11-18Context window 1.05MMax output 65.5KInput Text, image, video, audio, PDFOutput TextWeights ClosedData version 2026103001, as of 2026-10-30
Knowledge Work
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| APEX-Agents | 18.4% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: high |
| GDPval-AA (Elo) | 1195 | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 2 more pages | Effort: high |
| Vals Finance Agent v1.1 | 55.2% | Anthropic, Claude Sonnet 4.6checked 2026-10-15 | - |
| Vending-Bench 2 (final balance in USD) | $5478.16 | Z.ai, GLM-5 model cardchecked 2026-10-15 | - |
| Vals Finance Agent (version not stated) | 44.1% | Anthropic, Claude Opus 4.6checked 2026-10-15 | - |
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-bench Verified | 76.2% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 8 more pages | Effort: high |
| SWE-Bench Pro | 43.3% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: high |
| SWE-bench Multilingual | 65% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 2 more pages | - |
| Terminal-Bench 2.0 | 56.9% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 8 more pages | Effort: high; Harness: Terminus-2 |
| Terminal-Bench Hard | 39% | Z.ai, GLM-4.7 model cardchecked 2026-10-15 | - |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 91.9% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 8 more pages | Effort: high; Tools: none |
| AIME 2025 | 95% | DeepSeek, DeepSeek-V3.2 model cardchecked 2026-10-15; printed on 3 more pages | - |
| SciCode | 56% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 2 more pages | Effort: high |
| HMMT February 2025 | 97.3% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 3 more pages | - |
| IMO-AnswerBench | 83.3% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 3 more pages | - |
| AIME 2026 | 90.6% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 1 more page | - |
| HMMT November 2025 | 93.3% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 2 more pages | - |
| PolyMATH | 81.6% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| AA-LCR | 70.7% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 1 more page | - |
| LongBench v2 | 68.2% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 1 more page | - |
| MRCR v2 8-needle 128K | 77% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: high; Context: 128k (average) |
| MRCR v2 8-needle 1M (pointwise) | 26.3% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: high; Context: 1M (pointwise) |
Computer Use
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| ScreenSpot-Pro | 72.7% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
Multimodal Understanding
Cybersecurity
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| CyberGym | 39.9% | Moonshot AI, Kimi K2.5 model cardchecked 2026-10-15; printed on 1 more page | Effort: high |
| SecCodeBench | 62.4% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 37.5% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 8 more pages | Effort: high; Tools: none |
| MMLU-Pro | 89.8% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 2 more pages | - |
| Humanity's Last Exam (with tools) | 45.8% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 6 more pages | Effort: high; Tools: search (blocklist), code |
| MMLU-Redux | 95.9% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| C-Eval | 93.4% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| SuperGPQA | 74% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| ARC-AGI-2 | 31.1% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 3 more pages | Effort: high |
| Humanity's Last Exam (Verified) | 48% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
Code Generation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| LiveCodeBench v6 | 90.7% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 2 more pages | - |
| Codeforces rating | 2708 | DeepSeek, DeepSeek-V3.2 model cardchecked 2026-10-15 | - |
| OJBench (C++) | 68.5% | Moonshot AI, Kimi K2.5 model cardchecked 2026-10-15 | Effort: high |
| LiveCodeBench Pro (Elo) | 2439 | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: high |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Toolathlon | 36.4% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 2 more pages | - |
| τ²-bench Telecom | 98% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 3 more pages | Effort: high |
| MCP Atlas | 54.1% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 2 more pages | Effort: high |
| τ²-bench Retail | 85.3% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 3 more pages | Effort: high |
| MCP Atlas (public set) | 66.6% | Z.ai, GLM-5 model cardchecked 2026-10-15 | - |
| MCPMark | 53.9% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| τ²-bench (domain not stated) | 85.4% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 3 more pages | - |
| DeepPlanning | 23.3% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| VitaBench | 51.6% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| BFCL v4 | 72.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 59.2% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 6 more pages | Effort: high; Tools: search, python, browse |
| BrowseComp-ZH | 66.8% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 1 more page | Tools: enabled; search; Context: 256000 |
| WideSearch (metric not stated) | 68 | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | Tools: enabled; search; Context: 256000 |
| Seal-0 | 45.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 1 more page | Tools: enabled; search; Context: 256000 |
| FRAMES | 90.9% | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
| DeepSearchQA (accuracy) | 63.2% | Moonshot AI, Kimi K2.5 model cardchecked 2026-10-15 | Effort: high; Tools: search, code interpreter, browsing |
| WideSearch (item-F1) | 57% | Moonshot AI, Kimi K2.5 model cardchecked 2026-10-15 | Effort: high; Tools: search, code interpreter, browsing |
| FinSearchComp T2 and T3 | 49.9% | Moonshot AI, Kimi K2.5 model cardchecked 2026-10-15 | Effort: high; Tools: search, code interpreter, browsing |
| Research-Eval (Reka) | 55.9 | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
| X Browse | 26.5 | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
Instruction Following
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| IFBench | 70.4 | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15; printed on 1 more page | - |
| MultiChallenge | 64.2% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| IFEval | 93.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
Multilingual
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MMMLU | 91.8% | Google, Gemini 3.1 Pro model cardchecked 2026-10-15; printed on 4 more pages | Effort: high |
| MMLU-ProX | 87.7% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| WMT24++ (XCOMET-XXL) | 80.7 | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Global PIQA | 93.2% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| INCLUDE | 90.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| MaXIFE | 87.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| NOVA-63 | 56.7% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
Healthcare
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MedXpertQA MM | 76% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| PMC-VQA | 62.3% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| SLAKE | 81.3% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
More from Google
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.