Gemini 3 Pro vs Kimi K2.6
Wins 1 of 8 areas
Reasoning
Wins 5 of 8 areas
Coding · Agents · Math · Long documents · Following instructions
Kimi K2.6 is the stronger all-rounder.Gemini 3 Pro is better at reasoning.
Scores updated · 45 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software16Kimi K2.66 of 7 tests
- MathCompetition and research-level math03Kimi K2.63 of 3 tests
- AgentsCarrying out multi-step tasks on its own13Kimi K2.63 of 4 tests
- Long documentsFinding answers in very long texts01Kimi K2.61 of 1 test
- Following instructionsDoing exactly what it is asked01Kimi K2.61 of 1 test
- ReasoningHard problems that need careful thinking20Gemini 3 Pro2 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up22Even2 each
- Images and chartsUnderstanding pictures, charts and video11Even1 each · 1 tie
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.
Kimi K2.6 costs 65% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Kimi K2.6 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+51points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+27.1points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+15.3points ahead
Where Gemini 3 Pro pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+38points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+23.2points ahead
- Real work tasks from 44 professionsGDPVal+7.2points ahead
Every test, side by side
All 45 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingKimi K2.6
- SWE-bench ProKimi K2.6 by 15.343.358.6+15.3
- SWE-bench MultilingualKimi K2.6 by 11.76576.7+11.7
- LMArena · WebDevKimi K2.6 by 69 rating points14401509+69 rating
- SciCodeGemini 3 Pro by 4.656.151.5+4.6
- SWE-bench VerifiedKimi K2.6 by 476.280.2+4
- LiveCodeBench v6Kimi K2.6 by 2.287.489.6+2.2
- Terminal-Bench HardKimi K2.6 by 2.241.743.9+2.2
AgentsKimi K2.6
- BrowseCompKimi K2.6 by 27.159.286.3+27.1
- MCP AtlasKimi K2.6 by 1454.168.1+14
- AA ApexAgentsKimi K2.6 by 10.118.428.5+10.1
- GDPValGemini 3 Pro by 7.234.227+7.2
ReasoningGemini 3 Pro
- Humanity's Last ExamGemini 3 Pro by 2.239.737.5+2.2
- CritPtGemini 3 Pro by 1.19.18+1.1
- GPQA Diamondtie90.891.1tie
FactsEven
- AA-Omniscience · Non-hallucinationKimi K2.6 by 518.559.5+51
- SimpleQA VerifiedGemini 3 Pro by 3872.934.9+38
- AA-Omniscience · AccuracyGemini 3 Pro by 23.255.832.6+23.2
- Vectara HHEM hallucination ratelower is betterKimi K2.6 by 2.813.610.8+2.8
Images and chartsEven
- CharXiv (RQ)Kimi K2.6 by 5.381.486.7+5.3
- LMArena · VisionGemini 3 Pro by 23 rating points13051282+23 rating
- MMMU-Protie80.279.4tie
MathKimi K2.6
- HMMT Feb. 2026Kimi K2.6 by 6.386.492.7+6.3
- AIME 2026Kimi K2.6 by 4.791.796.4+4.7
- IMOAnswerBenchKimi K2.6 by 2.983.186+2.9
Other results19 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.
- AA Agentic IndexGemini 3 Pro by 29.95222.1+29.9
- GDPval-AA (Elo)Kimi K2.6 by 287 rating points11951482+287 rating
- WideSearch (item-f1)Kimi K2.6 by 23.85780.8+23.8
- DeepSearchQA (F1)Kimi K2.6 by 19.863.283+19.8
- Artificial Analysis Coding IndexKimi K2.6 by 15.346.561.8+15.3
- ToolathlonKimi K2.6 by 13.636.450+13.6
- Terminal-Bench 2.0Kimi K2.6 by 12.554.266.7+12.5
- AA-OmniscienceGemini 3 Pro by 1015.35.3+10
- τ²-Bench Telecom (AA run)Kimi K2.6 by 8.887.195.9+8.8
- HLE (with tools)Kimi K2.6 by 8.245.854+8.2
- MATH-VisionKimi K2.6 by 7.186.193.2+7.1
- frontiermath_tier_4_v1Gemini 3 Pro by 4.218.814.6+4.2
- MMLU-ProGemini 3 Pro by 390.187.1+3
- vectara_factual_consistencyKimi K2.6 by 2.886.489.2+2.8
- LiveCodeBenchGemini 3 Pro by 2.49289.6+2.4
- vectara_avg_summary_lengthKimi K2.6 by 14.8 rating points101.9116.7+14.8 rating
- LiveBenchGemini 3 Pro by 1.273.472.2+1.2
- AA IntelligenceGemini 3 Pro by 12827+1
- vectara_answer_ratetie99.499.7tie
Questions people ask
Which is better, Gemini 3 Pro or Kimi K2.6?
Kimi K2.6 wins five of the eight areas we test: coding, agents, math, long documents and following instructions. Gemini 3 Pro wins reasoning. They are level on facts and images and charts.
Which is better for coding?
Kimi K2.6. It wins 6 of the 7 coding tests both models report; Gemini 3 Pro wins 1.
Which is cheaper?
Gemini 3 Pro costs $2.00 per million input tokens and $12.00 per million output tokens; Kimi K2.6 costs $0.95 and $4.00. That makes Kimi K2.6 about 65% cheaper for the same work.
How do you compare the two?
We use the 45 benchmark tests both models have published scores on. The verdict counts the 26 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 19 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.