GPT-5.1 vs Grok 4
Wins 4 of 7 areas
Reasoning · Images and charts · Long documents · Following instructions
Wins 1 of 7 areas
Agents
GPT-5.1 wins more areas, narrowly.Grok 4 is better at agents.
Scores updated · 30 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- ReasoningHard problems that need careful thinking30GPT-5.13 of 4 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video20GPT-5.12 of 2 tests
- Long documentsFinding answers in very long texts10GPT-5.11 of 1 test
- Following instructionsDoing exactly what it is asked10GPT-5.11 of 1 test
- AgentsCarrying out multi-step tasks on its own01Grok 41 of 1 test
- CodingWriting and fixing software11Even1 each
- FactsGetting facts right instead of making them up11Even1 each · 1 tie
Agents, long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
GPT-5.1 costs 38% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GPT-5.1 pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+19.2points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+12.6points ahead
- Reasons across sets of long documentsAA-LCR+12points ahead
Where Grok 4 pulls ahead
- Real work tasks from 44 professionsGDPVal+8.1points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+2.8points ahead
- Code for real scientific research problemsSciCode+2.4points ahead
Every test, side by side
All 30 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingEven
- Terminal-Bench HardGPT-5.1 by 7.645.537.9+7.6
- SciCodeGrok 4 by 2.443.345.7+2.4
ReasoningGPT-5.1
- CritPtGPT-5.1 by 2.94.92+2.9
- Humanity's Last ExamGPT-5.1 by 1.828.526.7+1.8
- ARC-AGI-2GPT-5.1 by 1.717.615.9+1.7
- GPQA Diamondtie87.387.7tie
FactsEven
- AA-Omniscience · Non-hallucinationGPT-5.1 by 12.648.135.5+12.6
- AA-Omniscience · AccuracyGrok 4 by 2.837.740.5+2.8
- SimpleQA Verifiedtie4847.9tie
Images and chartsGPT-5.1
- MMMU-ProGPT-5.1 by 6.775.568.8+6.7
- LMArena · VisionGPT-5.1 by 40 rating points12501210+40 rating
Other results16 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- HarmBenchGPT-5.1 by 57.997.639.7+57.9
- AIR-Bench 2024GPT-5.1 by 41.886.244.4+41.8
- AA Agentic IndexGrok 4 by 19.921.641.5+19.9
- frontiermath_tier_4_v1GPT-5.1 by 10.412.52.1+10.4
- Artificial Analysis Coding IndexGPT-5.1 by 8.949.440.5+8.9
- simple_safety_testsGPT-5.1 by 7.699.892.2+7.6
- τ²-Bench Telecom (AA run)GPT-5.1 by 781.974.9+7
- matharena_visual_math_overallGPT-5.1 by 6.976.970+6.9
- ARC-AGI-1GPT-5.1 by 6.172.866.7+6.1
- bbqGrok 4 by 588.793.7+5
- anthropic_red_teamGPT-5.1 by 3.899.595.7+3.8
- HMMT Nov. 2025GPT-5.1 by 3.491.788.3+3.4
- AA-OmniscienceGPT-5.1 by 3.35.42.1+3.3
- AIME 2025GPT-5.1 by 2.39491.7+2.3
- AA IntelligenceGPT-5.1 by 2.224.722.5+2.2
- XSTestGPT-5.1 by 1.698.296.6+1.6
Questions people ask
Which is better, GPT-5.1 or Grok 4?
GPT-5.1 wins four of the seven areas where both have results: reasoning, images and charts, long documents and following instructions. Grok 4 wins agents. They are level on coding and facts.
Which is better for coding?
Neither. They win 1 coding test each of the 2 both models report.
Which is cheaper?
GPT-5.1 costs $1.25 per million input tokens and $10.00 per million output tokens; Grok 4 costs $3.00 and $15.00. That makes GPT-5.1 about 38% cheaper for the same work.
How do you compare the two?
We use the 30 benchmark tests both models have published scores on. The verdict counts the 14 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.