Gemini 3.1 Pro vs Grok 4.7
Full profiles: Gemini 3.1 Pro (Google) and Grok 4.7 (SpaceXAI): every benchmark, price and rank.
Updated · Data from 14 sources
As of 16 Nov 2026, Gemini 3.1 Pro and Grok 4.7 are neck and neck on quality (60.4 vs 57.4 on the BenchLeader Index, inside the ±4.3 margin), and Grok 4.7 is 1.5× cheaper at $3.00 vs $4.50 per 1M tokens blended. Choose Gemini 3.1 Pro for maths (66 vs 56), reasoning (65 vs 56), and knowledge (66 vs 60) and fast first answers (15.2 s vs 55.9 s), where it also streams ~1.5× faster; choose Grok 4.7 for coding (61 vs 59).
Key facts (as of 16 Nov 2026)
- Gemini 3.1 Pro scores 60.4 on the BenchLeader Index and Grok 4.7 scores 57.4 (BenchLeader, 16 Nov 2026).
- Gemini 3.1 Pro costs $4.50 per 1M tokens blended (list price, 16 Nov 2026).
- Grok 4.7 costs $3.00 per 1M tokens blended (list price, 16 Nov 2026).
- Gemini 3.1 Pro streams ~1.5× faster than Grok 4.7 (Artificial Analysis, 7-day median, 16 Nov 2026).
- Of the 40 benchmarks both models have, Grok 4.7 leads 28 (BenchLeader, 16 Nov 2026).
Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Grok 4.7 | 61 vs 59 category score; IOI 57.7% vs 51.8% | Medium |
| Agents & tool use | Either | 57 vs 57 category score; Terminal-Bench 2.1 (Vals) 70.8% vs 73.4% | Either |
| Maths | Gemini 3.1 Pro | 66 vs 56 category score; OTIS Mock AIME 95.6% vs 98.1% | High |
| Reasoning | Gemini 3.1 Pro | 65 vs 56 category score; GPQA Diamond 94.4% vs 92.7% | High |
| Knowledge | Gemini 3.1 Pro | 66 vs 60 category score; SimpleQA Verified 73.5% vs 56.0% | Medium |
| Chat quality (human preference) | – | Only Grok 4.7 has an Arena rating (63 category score) | Not comparable |
| Lowest cost | Grok 4.7 | $3.00 vs $4.50 per 1M blended; agentic coding session $0.144 vs $0.168 | – |
| Fastest first answer | Gemini 3.1 Pro | 15.2 s (Gemini 3.1 Pro (high)) vs 55.9 s (Grok 4.7 (xhigh)) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Gemini 3.1 Pro (high) and Grok 4.7 (xhigh) are neck and neck on quality (60.4 vs 57.4, within the ±4.3 margin).
- Gemini 3.1 Pro (high) is stronger in knowledge, maths, reasoning.
- Grok 4.7 (xhigh) is stronger in agents & tools, coding.
- Grok 4.7 (xhigh) is 1.5× cheaper ($3.00 vs $4.50 per 1M blended).
- Gemini 3.1 Pro (high) streams 1.6× faster (111 vs 68 tokens per second).
At a glance
| Spec | Gemini 3.1 Pro | Grok 4.7 |
|---|---|---|
| Developer | SpaceXAI | |
| Released | 19 Feb 2026 | 21 Sep 2026 |
| Context window (tokens) | 1,048,576 | 500,000 |
| Compared at | Gemini 3.1 Pro (high) | Grok 4.7 (xhigh) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $2.00 / $12.00 / $0.200 | $2.00 / $6.00 / $0.500 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Gemini 3.1 Pro (high) | Grok 4.7 (xhigh) | Leader |
|---|---|---|---|
| Reasoning | 64.7 | 55.6 | Gemini 3.1 Pro |
| Coding | 58.6 | 60.9 | Grok 4.7 |
| Agents & tools | 56.7 | 56.9 | level |
| Maths | 65.7 | 55.8 | Gemini 3.1 Pro |
| Knowledge | 65.6 | 59.9 | Gemini 3.1 Pro |
| Human preference | – | 62.5 | not comparable |
| Multimodal | – | 55.5 | not comparable |
| Long context | – | 62.9 | not comparable |
| Composite | – | 52.8 | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Instruction following, Long context, Multimodal, Human preference.
Where sources disagree
Across the 40 benchmarks both have, Gemini 3.1 Pro leads 12, Grok 4.7 leads 28. Every overall index that rates both puts Gemini 3.1 Pro ahead.
| Overall index | Gemini 3.1 Pro (high) | Grok 4.7 (xhigh) | Leader |
|---|---|---|---|
| BenchLeader Index | 60.4 | 57.4 | neck and neck |
| Arena Text | 1487† | 1443 | Gemini 3.1 Pro |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | Gemini 3.1 Pro (high) | Grok 4.7 (xhigh) |
|---|---|---|
| Blended price per 1M tokens | $4.50 | $3.00 |
| Input / output per 1M | $2.00 / $12.00 | $2.00 / $6.00 |
| Cached input per 1M | $0.200 | $0.500 |
| Price per task: chat reply | $0.0044 | $0.0026 |
| Price per task: summarize a 30-page report | $0.031 | $0.028 |
| Price per task: code edit | $0.030 | $0.021 |
| Price per task: agentic coding session | $0.168 | $0.144 |
| Price per task: structured extraction | $0.0064 | $0.0052 |
| Agentic coding session, input mostly cached | $0.087 | $0.076 |
| Cost per benchmark task (AA Index) | – | $7.77 |
| Output tokens per task (AA Index) | – | 105k |
| Of which reasoning | – | 69k |
| Time per task (AA Index) | – | 26 min |
| Output speed | 111 tok/s | 68 tok/s |
| Time to first answer | 15.2 s | 55.9 s |
| Time to a complete answer, last measured 13 Nov 2026 | 37.5 s | 145.5 s |
Each model is compared at its best-ranked reasoning-effort setting. Gemini 3.1 Pro (high) scores 60.4 and takes 15 s to its first answer token; Gemini 3.1 Pro scores 59.9 in 15 s. Grok 4.7 (xhigh) scores 57.4 and takes 56 s to its first answer token; Grok 4.7 (high) scores 56.9 in 35 s, Grok 4.7 (low) scores 55.4 in 10 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
212 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | Gemini 3.1 Pro (high) | Grok 4.7 (xhigh) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 60.4 | 57.4 |
| Agents & tools score | BenchLeader | 56.7 | 56.9 |
| Coding score | BenchLeader | 58.6 | 60.9 |
| Knowledge score | BenchLeader | 65.6 | 59.9 |
| Maths score | BenchLeader | 65.7 | 55.8 |
| Reasoning score | BenchLeader | 64.7 | 55.6 |
| Composite score | BenchLeader | – | 52.8 |
| Human preference score | BenchLeader | – | 62.5 |
| Long context score | BenchLeader | – | 62.9 |
| Multimodal score | BenchLeader | – | 55.5 |
| Blended price $/M | BenchLeader | $4.50 | $3.00 |
| Output speed | BenchLeader | 111 tok/s | 68 tok/s |
| Time to first answer | BenchLeader | 15.2 s | 55.9 s |
| Context window | BenchLeader | 1,048,576 tokens | 500,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 94.4% | 92.7% |
| Humanity's Last Exam | Scale AI / CAIS | 46.4%† | – |
| SimpleBench | SimpleBench | 79.6%† | – |
| Arena Hard Prompts | Arena (LMArena) | 1508† | 1466 |
| LiveBench 2.0 Reasoning | LiveBench | – | 82.7% |
| LiveBench 1.0 Reasoning | LiveBench | 84.0% | 82.7% |
| GPQA Diamond (AA) | Artificial Analysis | 94.1%† | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 47.0%† | 43.1% |
| HLE-Diamond (AA) | Artificial Analysis | 31.4%† | 23.2% |
| GPQA Diamond (Vals) | Vals AI | 95.5% | – |
| EnigmaEval | Scale AI SEAL | 36.8% | – |
| ARC-AGI-1 | ARC Prize | 98.0%† | 89.5% |
| ARC-AGI-2 | ARC Prize | 77.1%† | 61.4% |
| ARC-AGI-3 | ARC Prize | 0.4%† | 7.0% |
| CritPt | Artificial Analysis | 17.7%† | 17.7% |
| CritPt v2 hard subset (AA) | Artificial Analysis | – | 33.5% |
| MysteryMechanism | Vals AI | 20.7% | 25.2% |
| HLE-Diamond (Scale) | Scale AI SEAL | – | 23.4%† |
| Arena Expert | Arena (LMArena) | 1509† | 1478 |
| Chess Puzzles | Epoch AI Benchmarking Hub | 51.0% | 41.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 34.0% | 29.0% |
| CL-bench | Epoch AI Benchmarking Hub | 20.8%† | – |
| CL-bench Life | Epoch AI Benchmarking Hub | 16.9%† | – |
| LMCA | Epoch AI Benchmarking Hub | 53.8% | 49.4% |
| DTBench | Epoch AI Benchmarking Hub | 95.5% | 96.0% |
Show all 212 benchmarks
| Metric | Publisher | Gemini 3.1 Pro (high) | Grok 4.7 (xhigh) |
|---|---|---|---|
| Reasoning | |||
| ForecastBench | ForecastBench | 65.2% | – |
| Coding | |||
| SciCode | SciCode | 58.7%† | 57.4% |
| WeirdML | WeirdML | 72.1%† | – |
| FrontierCode | Cognition | – | 47.6%† |
| GSO-Bench | GSO-Bench | 22.6%† | – |
| Arena Coding | Arena (LMArena) | 1520† | 1493 |
| Arena WebDev | Arena (LMArena) | 1446† | 1641 |
| LiveBench 2.0 Coding | LiveBench | – | 73.1% |
| LiveBench 1.0 Coding | LiveBench | 76.5% | 77.2% |
| LiveCodeBench | Vals AI | 88.5% | – |
| IOI | Vals AI | 51.8% | 57.7% |
| SWE-bench (Vals) | Vals AI | 78.8% | – |
| SWE-Bench Pro | Scale AI SEAL | 46.1%† | – |
| Code Migration | Vals AI | 17.3% | 44.8% |
| Code Migration v1.1 | Vals AI | – | 53.2% |
| ProgramBench | Vals AI | 0.0% | 0.5% |
| SWE-bench Multimodal | Vals AI | – | 42.9% |
| Vibe Code Bench 1-100 | Vals AI | 6.7% | – |
| Vibe Code Bench v1.1 | Vals AI | 32.0% | 86.2% |
| Vibe Code Bench 1-100 v1.1 | Vals AI | – | 26.4% |
| SWE-Bench Pro (private) | Scale AI SEAL | 32.2%† | – |
| SWE Atlas v1.0: Codebase QnA | Scale AI SEAL | 13.5%† | – |
| SWE Atlas v1.0: Refactoring | Scale AI SEAL | 33.8%† | – |
| SWE Atlas v1.0: Test Writing | Scale AI SEAL | 29.8%† | – |
| SWE Atlas v1.1 | Scale AI SEAL | – | 62.1%† |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | – | 73.4%† |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | – | 64.4%† |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | – | 48.6%† |
| MirrorCode | Epoch AI Benchmarking Hub | 8.9% | 54.4% |
| DeepSWE v1.1 | Epoch AI Benchmarking Hub | 11.7%† | – |
| CursorBench | Cursor | – | 46.3% |
| ALE-Bench | Epoch AI Benchmarking Hub | 1160.6† | 1545.8 |
| AlgoTune | AlgoTune | 2.0† | – |
| FrontierSWE | Epoch AI Benchmarking Hub | – | 29.5% |
| WeirdML v3 | WeirdML | – | 8.4% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 4.0%† | 25.8% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 73.8%† | – |
| Startrise Frontend | Startrise | – | 75.9%† |
| GameBench 2 Overall | DuelLab GameBench 2 | – | 63.6† |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | – | 71.2 |
| Design Arena Website | Design Arena | 1226† | 1244† |
| FrontierCode 1.1 (AA) | Artificial Analysis | – | 33.8% |
| FrontierCode 2.0 (AA) | Artificial Analysis | – | 28.7% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 80.2%† | 38.8% |
| APEX-Agents | Mercor | 35.1%† | 54.6%† |
| Arena Agent | Arena (LMArena) | -4.5† | 2.6 |
| LiveBench Long Agentic | LiveBench | – | 45.3% |
| LiveBench Browser | LiveBench | – | 49.3% |
| LiveBench Agentic Coding | LiveBench | 44.1% | 54.0% |
| Terminal-Bench Hard | Artificial Analysis | 53.8%† | – |
| τ²-Bench Telecom (AA) | Artificial Analysis | 95.6%† | – |
| Terminal-Bench 2.1 (Vals) | Vals AI | 70.8% | 73.4% |
| MCP Atlas | Scale AI SEAL | 78.2% | – |
| HiL-Bench | Scale AI SEAL | 35.3%† | – |
| τ³-bench Banking | τ²-bench | 26.0% | – |
| GDPval-AA v2.1, normalised score | Artificial Analysis | 14.6%† | 60.7% |
| GDPval-AA v2.1 | Artificial Analysis | 799† | 1717 |
| τ³-Banking (AA) | Artificial Analysis | 21.4%† | – |
| ITBench SRE (AA) | Artificial Analysis | 30.3%† | 42.1% |
| Analyst Agent (AA) | Artificial Analysis | 41.3%† | 52.5% |
| APEX-Agents (AA) | Artificial Analysis | 32.0%† | – |
| CyberBench | Vals AI | – | 69.5% |
| Finance Agent v2 | Vals AI | 43.0% | 52.3% |
| Finance Agent v2.1 | Vals AI | – | 60.6% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | – | 18.3% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 0.0% | 12.5% |
| Legal Research Bench | Vals AI | 20.7% | 47.1% |
| Robo-bench | Vals AI | – | 0.2% |
| SREBench | Vals AI | 0.0% | 3.8% |
| Tax Agent Bench | Vals AI | 9.3% | 35.8% |
| Terminal-Bench 2.0 (Vals) | Vals AI | 67.4% | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | 2.5% | 28.8% |
| Terminal-Bench Science | Vals AI | 1.4% | 10.0% |
| Arena Search | Arena (LMArena) | 1209† | – |
| Arena Agent Task Completion | Arena (LMArena) | -11.3† | 5.1 |
| BALROG | BALROG | 57.0%† | – |
| DeepResearch Bench | DeepResearch Bench | 47.8% | – |
| PostTrainBench | Epoch AI Benchmarking Hub | 18.1%† | – |
| ExploitBench | Epoch AI Benchmarking Hub | 26.1%† | – |
| METR Time Horizons | METR | 77.0%† | – |
| GBAEval | Epoch AI Benchmarking Hub | 0.8%† | – |
| Vending-Bench 2 | Andon Labs | 911.2† | 10536.8† |
| AutomationBench | Artificial Analysis | 35.4%† | 65.6% |
| GDP.pdf (AA) | Artificial Analysis | 17.8%† | 20.0% |
| Harvey LAB-AA v1.1 | Harvey | 0.0%† | 9.4% |
| EnterpriseOps-Gym | Artificial Analysis | 42.2%† | – |
| AA-Briefcase v1.1 | Artificial Analysis | 457† | 1636 |
| GDPx | GDPx by SemiAnalysis | – | 3.9% |
| Design Arena Agentic | Design Arena | 1088† | 1277† |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | – | 11.4% |
| CWE-Bench (AA) | Artificial Analysis | – | 68.3% |
| DeepSec-Bench (AA) | Artificial Analysis | – | 26.9% |
| CyberGym end-to-end (AA) | Artificial Analysis | – | 74.0% |
| MKA (AA) | Artificial Analysis | – | 40.1% |
| BrowseComp | OpenAI | 31.2%† | – |
| DeepSearchQA | 60.2%† | – | |
| FACTS Search | 90.2%† | 85.6%† | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 59.6%† | 53.0% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 26.8%† | 17.1% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 95.6% | 98.1% |
| ProofBench | Vals AI | 26.0%† | 26.0%† |
| LiveBench Mathematics | LiveBench | 91.0% | 95.7% |
| AIME (Vals) | Vals AI | 98.1% | – |
| AIME 2026 | MathArena | 98.3% | – |
| HMMT February 2026 | MathArena | 94.7% | – |
| USAMO 2026 | MathArena | 74.4% | – |
| MathArena Apex | MathArena | 60.9% | – |
| Arena Maths | Arena (LMArena) | 1488† | 1447 |
| FrontierMath Erdős | Epoch AI Benchmarking Hub | – | 0.0% |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 73.5% | 56.0% |
| LiveBench 2.0 Data Analysis | LiveBench | – | 76.9% |
| LiveBench Research | LiveBench | – | 89.5% |
| LiveBench 1.0 Data Analysis | LiveBench | 78.5% | 76.9% |
| AA-Omniscience v1.1 | Artificial Analysis | – | 31.2 |
| MMLU-Pro | Vals AI | 91.0% | – |
| LegalBench | Vals AI | 87.4% | 84.4% |
| CorpFin | Vals AI | 64.5% | – |
| TaxEval | Vals AI | 72.9% | – |
| MedQA | Vals AI | 96.4% | – |
| PRBench Finance | Scale AI SEAL | 41.9%† | – |
| PRBench Legal | Scale AI SEAL | 44.0%† | 47.6% |
| MultiNRC | Scale AI SEAL | 64.7%† | – |
| BioMysteryBench | Vals AI | 60.0% | 69.3% |
| CaseLaw v2 | Vals AI | 64.8% | – |
| Excel Modeling Benchmark | Vals AI | 52.6% | 67.0% |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 71.5% |
| MedCode | Vals AI | 59.1% | 49.5% |
| MedScribe | Vals AI | 76.1% | 89.4% |
| Public Benefits Bench v1.1 | Vals AI | 53.8% | 65.6% |
| Access to Justice Benchmark | Vals AI | – | 73.7% |
| Financial Coaching Bench | Vals AI | 66.7% | 74.1% |
| Immigration Bench | Vals AI | – | 68.0% |
| DrugDiscoveryBench | Scale AI SEAL | 41.9% | – |
| EBR-bench | Epoch AI Benchmarking Hub | 14.3%† | – |
| GDP.pdf | Epoch AI Benchmarking Hub | 17.0% | 22.8% |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | – | 42.7% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | – | 80.0% |
| Global-MMLU-Lite (AA) | Artificial Analysis | 93.2%† | – |
| HealthBench Professional (AA) | OpenAI | – | 55.8% |
| FACTS Parametric | 79.4%† | 54.3%† | |
| Instruction following | |||
| LiveBench Language | LiveBench | 85.4% | 80.1% |
| LiveBench Instruction Following | LiveBench | 79.1% | 75.3% |
| IFBench | Artificial Analysis | 77.1%† | – |
| MultiChallenge | Scale AI SEAL | 71.4%† | – |
| TutorBench | Scale AI SEAL | 53.0%† | – |
| Arena Instruction Following | Arena (LMArena) | 1480† | 1435 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1487† | 1443 |
| EQ-Bench 4 | EQ-Bench | 1132† | – |
| EQ-Bench Creative Writing v3 | EQ-Bench | 1491† | 2012† |
| Arena Creative Writing | Arena (LMArena) | 1482† | 1421 |
| Arena Multi-turn | Arena (LMArena) | 1495† | 1435 |
| Arena Longer Queries | Arena (LMArena) | 1499† | 1454 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1296† | 1230 |
| LiveBench Design | LiveBench | – | 75.0% |
| MMMU-Pro | Artificial Analysis | 82.4%† | 78.2%† |
| MMMU-Pro (Vals) | Vals AI | 88.2% | – |
| MortgageTax | Vals AI | 69.4% | – |
| SAGE | Vals AI | 83.2% | 88.7% |
| VTB | Scale AI SEAL | 29.0%† | – |
| GeoBench | GeoBench | 84.0%† | 61.0%† |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 26.5%† | 32.5%† |
| Furniture Assembly | Epoch AI Benchmarking Hub | 26.7% | 20.8% |
| VIALS-AA | Handshake AI | – | 37.9% |
| Code2Video-AA | Artificial Analysis | – | 1127 |
| AA-3D | Artificial Analysis | – | 992 |
| FACTS Multimodal | 50.4%† | 40.9%† | |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 82.0%† | 76.7% |
| Arena Document | Arena (LMArena) | 1459† | – |
| MLCR | Artificial Analysis | 11.1%† | 20.0% |
| FACTS Grounding | 56.6%† | 68.6%† | |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 156.2† | 156.2† |
| LiveBench 2.0 | LiveBench | – | 67.9% |
| LiveBench 1.0 | LiveBench | 77.0% | 77.4% |
| AA Intelligence Index v5.0.1 | Artificial Analysis | – | 35.5 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 29.7† | 46.5 |
| AA Progress Index v1 | Artificial Analysis | 296.1† | 322.7 |
| Vals Index | Vals AI | – | 58.6 |
| Vals Finance Index | Vals AI | – | 58.8% |
| Vals RSI Index | Vals AI | – | 20.2% |
| Vals Science Index | Vals AI | 27.4% | 34.8% |
| Vals Multimodal Index | Vals AI | 56.1% | – |
| Vals Router Index | Vals AI | – | 58.6%† |
| Vals Legal Index | Vals AI | – | 72.7% |
| AA Cyber Index | Artificial Analysis | – | 56.4 |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 29.8%† | – |
| MASK | Scale AI SEAL | 42.4%† | – |
| DistressBench | Scale AI SEAL | 68.8%† | 52.4%† |
Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- Grok 4.7 (xhigh): BenchLeader Index 58.0 → 57.4.
- Grok 4.7 (xhigh): 8 new results (LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…).
- Gemini 3.1 Pro (high): 4 new results (AIME 2026, HMMT February 2026, USAMO 2026…).
- Grok 4.7 (xhigh): 1 new result (MirrorCode).
- Grok 4.7 (xhigh): 1 new result (ALE-Bench).
- Grok 4.7 (xhigh): 2 score revisions.
- Grok 4.7 (xhigh): 2 score revisions.
- Grok 4.7 (xhigh): 2 score revisions.
- Grok 4.7 (xhigh): 1 score revision.
- Grok 4.7 (xhigh): BenchLeader Index 59.5 → 58.0.
Related comparisons
- Gemini 3.1 Pro vs Gemini 4 Carbon
- Gemini 3.1 Pro vs Claude Opus 5.5
- Grok 4.7 vs Claude Opus 5.5
- Grok 4.7 vs GPT-6 Astra
- Grok 4.7 vs Claude Fable 5.1
- Grok 4.7 vs GPT-6.1 Sol
- Gemini 3.1 Pro vs Grok 4.6
- Gemini 3.1 Pro vs GPT-6.1 Sol
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Gemini 3.1 Pro vs Grok 4.7: questions
- Is Gemini 3.1 Pro better than Grok 4.7?
- Gemini 3.1 Pro (high) and Grok 4.7 (xhigh) are neck and neck on quality (60.4 vs 57.4, within the ±4.3 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
- Gemini 3.1 Pro vs Grok 4.7: which is cheaper, and will that change?
- Grok 4.7 is cheaper today: $3.00 against $4.50 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Gemini 3.1 Pro or Grok 4.7 better for coding?
- Grok 4.7 scores higher in coding (61 vs 59 on the category index, where 50 is average). On IOI, the top shared coding benchmark, it is 51.8% for Gemini 3.1 Pro against 57.7% for Grok 4.7.
- Which is faster, Gemini 3.1 Pro or Grok 4.7?
- Gemini 3.1 Pro streams ~1.5× faster (111 vs 68 output tokens per second, Artificial Analysis 7-day median).
- Is Gemini 3.1 Pro or Grok 4.7 better for agents and tool use?
- They are level in agentic tasks (57 vs 57 on the category index). On Terminal-Bench 2.1 (Vals), the top shared agents & tools benchmark, it is 70.8% for Gemini 3.1 Pro against 73.4% for Grok 4.7.
More questions (4)
- Which is better for long documents?
- Gemini 3.1 Pro accepts more context: 1,048,576 against 500,000 tokens for Grok 4.7.
- Which has lower latency?
- Gemini 3.1 Pro answers first: 15.2 s against 55.9 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Do independent benchmarks agree on Gemini 3.1 Pro vs Grok 4.7?
- Across the 40 benchmarks both have, Gemini 3.1 Pro leads 12, Grok 4.7 leads 28. Every overall index that rates both puts Gemini 3.1 Pro ahead.
- When was Grok 4.7 released?
- Grok 4.7 was released on 21 Sep 2026.