Claude Sonnet 4.5 vs GLM-4.6
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
Wins 1 of 7 areas
Math
Claude Sonnet 4.5 is the stronger all-rounder.GLM-4.6 is cheaper and better at math.
Scores updated · 47 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.
- CodingWriting and fixing software71Claude Sonnet 4.57 of 8 tests
- ReasoningHard problems that need careful thinking20Claude Sonnet 4.52 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own21Claude Sonnet 4.52 of 3 tests
- FactsGetting facts right instead of making them up21Claude Sonnet 4.52 of 3 tests
- Following instructionsDoing exactly what it is asked10Claude Sonnet 4.51 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts10Claude Sonnet 4.51 of 1 test
- MathCompetition and research-level math01GLM-4.61 of 1 test
Math and long documents 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.
GLM-4.6 costs 85% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Claude Sonnet 4.5 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+45points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+33.9points ahead
- Reasons across sets of long documentsAA-LCR+18.3points ahead
Where GLM-4.6 pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+21points ahead
- Recent programming contest problemsLiveCodeBench v6+18.8points ahead
- Olympiad math problems with short answersIMOAnswerBench+7.6points ahead
Every test, side by side
All 47 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.5
- SWE-bench ProClaude Sonnet 4.5 by 33.943.69.7+33.9
- LiveCodeBench v6GLM-4.6 by 18.86482.8+18.8
- SWE-bench MultilingualClaude Sonnet 4.5 by 13.26753.8+13.2
- Terminal-Bench HardClaude Sonnet 4.5 by 10.635.625+10.6
- SWE-bench VerifiedClaude Sonnet 4.5 by 9.277.268+9.2
- SciCodeClaude Sonnet 4.5 by 7.345.738.4+7.3
- Terminal-Bench 2.1Claude Sonnet 4.5 by 6.455.849.4+6.4
- LMArena · WebDevClaude Sonnet 4.5 by 53 rating points13941341+53 rating
AgentsClaude Sonnet 4.5
- BrowseCompGLM-4.6 by 2124.145.1+21
- τ-Bench V3 · BankingClaude Sonnet 4.5 by 11.124.513.4+11.1
- GDPValClaude Sonnet 4.5 by 7.520.513+7.5
ReasoningClaude Sonnet 4.5
- GPQA DiamondClaude Sonnet 4.5 by 5.483.478+5.4
- Humanity's Last ExamClaude Sonnet 4.5 by 3.317.814.5+3.3
- CritPttie1.11.1tie
FactsClaude Sonnet 4.5
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.5 by 4550.95.9+45
- AA-Omniscience · AccuracyClaude Sonnet 4.5 by 632.926.9+6
- Vectara HHEM hallucination ratelower is betterGLM-4.6 by 2.5129.5+2.5
Long documentsClaude Sonnet 4.5
- AA-LCRClaude Sonnet 4.5 by 18.372.354+18.3
Following instructionsClaude Sonnet 4.5
- IFBenchClaude Sonnet 4.5 by 13.957.343.4+13.9
- Multi-Challengetie55.354.9tie
Other results26 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-OmniscienceClaude Sonnet 4.5 by 41.8-0.1-41.9+41.8
- FinSearchComp-globalClaude Sonnet 4.5 by 31.660.829.2+31.6
- Terminal-Bench 2.0Claude Sonnet 4.5 by 25.55024.5+25.5
- τ²-BenchClaude Sonnet 4.5 by 22.89875.2+22.8
- HMMT Feb. 2025GLM-4.6 by 21.767.589.2+21.7
- Multi-SWE-BenchClaude Sonnet 4.5 by 14.344.330+14.3
- HMMT 2025GLM-4.6 by 14.174.688.7+14.1
- HMMT Nov. 2025GLM-4.6 by 1081.791.7+10
- Terminal-BenchClaude Sonnet 4.5 by 9.55040.5+9.5
- LiveCodeBenchGLM-4.6 by 8.57179.5+8.5
- τ²-Bench Telecom (AA run)Claude Sonnet 4.5 by 7.678.170.5+7.6
- BrowseComp-ZHGLM-4.6 by 7.142.449.5+7.1
- AIME 2025GLM-4.6 by 6.98793.9+6.9
- Artificial Analysis Coding IndexClaude Sonnet 4.5 by 6.352.145.8+6.3
- theagentcompanyClaude Sonnet 4.5 by 64135+6
- vectara_avg_summary_lengthClaude Sonnet 4.5 by 50.6 rating points127.877.2+50.6 rating
- MMLU-ProClaude Sonnet 4.5 by 4.387.583.2+4.3
- xbench-DeepSearchGLM-4.6 by 46670+4
- vectara_factual_consistencyGLM-4.6 by 2.58890.5+2.5
- AA IntelligenceClaude Sonnet 4.5 by 2.220.718.5+2.2
- frontiermath_tier_4_v1Claude Sonnet 4.5 by 2.14.22.1+2.1
- ArtifactsBenchClaude Sonnet 4.5 by 1.761.559.8+1.7
- HLE (with tools)Claude Sonnet 4.5 by 1.63230.4+1.6
- AA Agentic IndexGLM-4.6 by 1.117.518.6+1.1
- vectara_answer_rateClaude Sonnet 4.5 by 1.195.694.5+1.1
- GAIA (text only)tie71.271.9tie
Questions people ask
Which is better, Claude Sonnet 4.5 or GLM-4.6?
Claude Sonnet 4.5 wins six of the seven areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. GLM-4.6 wins math, and costs 85% less.
Which is better for coding?
Claude Sonnet 4.5. It wins 7 of the 8 coding tests both models report; GLM-4.6 wins 1.
Which is cheaper?
Claude Sonnet 4.5 costs $3.00 per million input tokens and $15.00 per million output tokens; GLM-4.6 costs $0.57 and $2.20. That makes GLM-4.6 about 85% cheaper for the same work.
How do you compare the two?
We use the 47 benchmark tests both models have published scores on. The verdict counts the 21 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 26 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.