GLM-5.1 vs Step 3.5 Flash
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
Wins 1 of 7 areas
Math
GLM-5.1 is the stronger all-rounder.Step 3.5 Flash is cheaper and better at math.
Scores updated · 25 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 software30GLM-5.13 of 4 tests · 1 tie
- ReasoningHard problems that need careful thinking30GLM-5.13 of 3 tests
- AgentsCarrying out multi-step tasks on its own20GLM-5.12 of 2 tests
- FactsGetting facts right instead of making them up10GLM-5.11 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts10GLM-5.11 of 1 test
- Following instructionsDoing exactly what it is asked10GLM-5.11 of 1 test
- MathCompetition and research-level math03Step 3.5 Flash3 of 3 tests
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.
Step 3.5 Flash costs 93% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GLM-5.1 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+55.8points ahead
- Reasons across sets of long documentsAA-LCR+23.7points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+15.9points ahead
Where Step 3.5 Flash pulls ahead
- Harvard-MIT high-school math contest problemsHMMT Feb. 2026+3.8points ahead
- Olympiad math problems with short answersIMOAnswerBench+1.6points ahead
- US invitational high-school math exam problemsAIME 2026+1.4points ahead
Every test, side by side
All 25 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGLM-5.1
- Terminal-Bench HardGLM-5.1 by 15.943.227.3+15.9
- SciCodeGLM-5.1 by 4.444.840.4+4.4
- SWE-bench VerifiedGLM-5.1 by 1.876.274.4+1.8
- LiveCodeBench v6tie85.786.4tie
ReasoningGLM-5.1
- Humanity's Last ExamGLM-5.1 by 930.121.1+9
- GPQA DiamondGLM-5.1 by 3.786.883.1+3.7
- CritPtGLM-5.1 by 2.14.62.5+2.1
FactsGLM-5.1
- AA-Omniscience · Non-hallucinationGLM-5.1 by 55.870.114.3+55.8
- AA-Omniscience · Accuracytie23.723.6tie
MathStep 3.5 Flash
- HMMT Feb. 2026Step 3.5 Flash by 3.882.686.4+3.8
- IMOAnswerBenchStep 3.5 Flash by 1.683.885.4+1.6
- AIME 2026Step 3.5 Flash by 1.495.396.7+1.4
Other results9 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-OmniscienceGLM-5.1 by 42.70.8-41.8+42.7
- AA Agentic IndexStep 3.5 Flash by 26.825.252+26.8
- Artificial Analysis Coding IndexGLM-5.1 by 24.255.831.6+24.2
- Terminal-Bench 2.0GLM-5.1 by 186951+18
- browsecomp_with_context_managerGLM-5.1 by 10.379.369+10.3
- AA IntelligenceGLM-5.1 by 9.526.116.6+9.5
- τ²-Bench Telecom (AA run)GLM-5.1 by 3.397.794.4+3.3
- MMLU-ProGLM-5.1 by 1.68684.4+1.6
- HMMT Nov. 2025tie9494tie
Questions people ask
Which is better, GLM-5.1 or Step 3.5 Flash?
GLM-5.1 wins six of the seven areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. Step 3.5 Flash wins math, and costs 93% less.
Which is better for coding?
GLM-5.1. It wins 3 of the 4 coding tests both models report; Step 3.5 Flash wins none, and 1 is a tie.
Which is cheaper?
GLM-5.1 costs $1.38 per million input tokens and $4.40 per million output tokens; Step 3.5 Flash costs $0.10 and $0.30. That makes Step 3.5 Flash about 93% cheaper for the same work.
How do you compare the two?
We use the 25 benchmark tests both models have published scores on. The verdict counts the 16 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 9 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.