GLM-5.3 vs Qwen3.8 27B
Wins 5 of 7 areas
Coding · Agents · Reasoning · Facts · Math
Wins 2 of 7 areas
Long documents · Following instructions
GLM-5.3 is the stronger all-rounder.Qwen3.8 27B is cheaper and better at long documents.
Scores updated · 33 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 thinking50GLM-5.35 of 5 tests
- CodingWriting and fixing software40GLM-5.34 of 5 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own30GLM-5.33 of 3 tests
- FactsGetting facts right instead of making them up10GLM-5.31 of 2 tests · 1 tie
- MathCompetition and research-level math10GLM-5.31 of 1 test
- Long documentsFinding answers in very long texts01Qwen3.8 27B1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.8 27B1 of 1 test
Math, 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.
Qwen3.8 27B costs 31% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where GLM-5.3 pulls ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+36.3points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+18.3points ahead
- Research-level physics problemsCritPt+13.7points ahead
Where Qwen3.8 27B pulls ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+3.4points ahead
- Reasons across sets of long documentsAA-LCR+2.3points ahead
Every test, side by side
All 33 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGLM-5.3
- SciCodeGLM-5.3 by 12.45946.6+12.4
- Terminal-Bench 2.1GLM-5.3 by 4.183.979.8+4.1
- LiveBench · CodingGLM-5.3 by 3.37975.7+3.3
- LMArena · WebDevGLM-5.3 by 33 rating points16231590+33 rating
- LiveBench · Agentic Codingtie60.961.4tie
AgentsGLM-5.3
- Terminal-Bench 4.0GLM-5.3 by 36.341.95.6+36.3
- GDPValGLM-5.3 by 11.557.546+11.5
- τ-Bench V3 · BankingGLM-5.3 by 2.350.348+2.3
ReasoningGLM-5.3
- CritPtGLM-5.3 by 13.719.15.4+13.7
- Humanity's Last ExamGLM-5.3 by 8.442.333.9+8.4
- SimpleBenchGLM-5.3 by 666.260.2+6
- LiveBench · ReasoningGLM-5.3 by 5.885.880+5.8
- GPQA DiamondGLM-5.3 by 1.291.790.5+1.2
FactsGLM-5.3
- AA-Omniscience · AccuracyGLM-5.3 by 18.333.915.6+18.3
- AA-Omniscience · Non-hallucinationtie70.469.7tie
Following instructionsQwen3.8 27B
- LiveBench · Instruction FollowingQwen3.8 27B by 3.469.372.7+3.4
Other results15 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.
- Terminal-Bench 4.0GLM-5.3 by 36.241.85.6+36.2
- CritPt v2 HardGLM-5.3 by 29.533.54+29.5
- DeepSWE 1.1GLM-5.3 by 24.766.942.2+24.7
- AA-OmniscienceGLM-5.3 by 24.314.3-10+24.3
- NL2RepoGLM-5.3 by 15.75842.3+15.7
- Agents' Last ExamQwen3.8 27B by 14.428.542.9+14.4
- AA IntelligenceGLM-5.3 by 11.144.833.7+11.1
- Terminal-Bench-Science 0.1GLM-5.3 by 9.59.50+9.5
- AA Agentic IndexGLM-5.3 by 6.953.446.5+6.9
- Artificial Analysis Coding IndexGLM-5.3 by 6.774.868.1+6.7
- livebench_data_analysisQwen3.8 27B by 6.470.276.6+6.4
- Toolathlon VerifiedGLM-5.3 by 5.97367.1+5.9
- livebench_languageGLM-5.3 by 5.579.974.3+5.5
- GDP (Surge AI)Qwen3.8 27B by 5.211.216.4+5.2
- HLE-DiamondGLM-5.3 by 2.617.915.3+2.6
Questions people ask
Which is better, GLM-5.3 or Qwen3.8 27B?
GLM-5.3 wins five of the seven areas where both have results: coding, agents, reasoning, facts and math. Qwen3.8 27B wins long documents and following instructions, and costs 31% less.
Which is better for coding?
GLM-5.3. It wins 4 of the 5 coding tests both models report; Qwen3.8 27B wins none, and 1 is a tie.
Which is cheaper?
GLM-5.3 costs $1.13 per million input tokens and $3.96 per million output tokens; Qwen3.8 27B costs $0.50 and $3.00. That makes Qwen3.8 27B about 31% cheaper for the same work.
How do you compare the two?
We use the 33 benchmark tests both models have published scores on. The verdict counts the 18 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 15 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.