GLM-5.3
- Input
- $1.40/M tok
- Output
- $4.40/M tok
- Blended
- $2.15/M tok · 3:1
- Speed
- 80tok/s
- Context
- 1Mtokens
- Cost per task
- $5.0/task · AA
Open weights (744B-A40B).
Rankings
Coding
Verified#17 / 137rank by quality74.1±4.5quality / 1004/7 benchmarks measuredWriting & chat
Verified#7 / 162rank by quality82.0±4.3quality / 1003/5 benchmarks measuredPrice frontierHard reasoning
Verified#24 / 218rank by quality76.4±3.7quality / 1003/4 benchmarks measuredAgents
Verified#6 / 187rank by quality85.3±5.9quality / 1003/7 benchmarks measuredPrice frontier
Self-hostingEstimated
How we estimate →What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.
- Hardware
- 16× H200 141GB
- FP8 weights
- Throughput
- ~29,333 tok/s
- many requests batched
- Price
- $0.28–2.33 /M tok
- busy → light use
- On your own machine
- Mac Studio M5 Ultra 512GB (int4, ~42 tok/s single-stream)
- single consumer GPU or Mac
Assumptions (5)
- FP8 weights (744B params, 40B active per token) + 40% KV-cache headroom ≈ 1042 GB
- 16× H200 141GB at $3.59–$7.91/GPU-hour on-demand (2026-09-24)
- ~29,333 output tok/s aggregate at batch 1241 (bandwidth-bound); MoE compute scales with active params
- Blended 3:1 input:output; prefill ~158,963 tok/s
- Low = 75% utilization at the low GPU price; high = 20% at the high price
Benchmark scores
~ italic, dashed = no published score yet, estimated from related benchmarks.
Coding
SWE-bench VerifiedEstimated
~#2 of 21 (estimated)~88.9%
No published score — estimated from related benchmarks
SWE-bench Pro (Public)Estimated
~#6 of 24 (estimated)~72.0%
- #1Claude Fable 5.293.0%
- #2Claude Opus 5.589.9%
- #5Claude Opus 579.2%
- ~#6GLM-5.3Estimated72.0%
- #6Intern-S2-397B68.5%
No published score — estimated from related benchmarks
DeepSWE v1.1Verified
#29 of 4561.4%
- #1Gemini 4 Argon78.8%
- #2DeepSeek V4.1 Flash 101777.0%
- #28Mistral Large 461.7%
- #29GLM-5.361.4%
- #30SWE-2 Flash60.1%
artificialanalysis.ai · 2026-09-24
Terminal-Bench 4.0Verified
#14 of 10141.9%
- #1Kimi K3.1 Pro67.2%
- #2Claude Sonnet 5.563.6%
- #13GPT-6 Sol43.9%
- #14GLM-5.341.9%
- #15Qwen3.8 Max (0902)38.9%
artificialanalysis.ai · 2026-09-24
LiveCodeBenchEstimated
~#10 of 28 (estimated)~87.3%
- #1Qwen3.8-Omni-Flash92.6%
- #2Qwen3.8 Flash-Next91.9%
- #9Nemotron 3.5 Super VL NVFP487.4%
- ~#10GLM-5.3Estimated87.3%
- #10Ternary Bonsai 27B82.8%
No published score — estimated from related benchmarks
Writing & chat
IFBenchEstimated
~#15 of 140 (estimated)~78.9%
- #1MiniMax-M382.9%
- #2Qwen3.8 2.4T A95B82.8%
- #14Qwen3.8 27B NVFP478.9%
- ~#15GLM-5.3Estimated78.9%
- #15Qwen3.5 397B A17B78.8%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Verified
#12 of 561463 Elo
- #1Gemini 4 ArgonFew votes1512
- #2Claude Opus 5.51497
- #11GPT-6.1 Sol1469
- #12GLM-5.31463
- #12GLM-5.3 Flash1463
lmarena.ai · 2026-10-02
LMArena Text (Elo)Verified
#12 of 561480 Elo
- #1Gemini 4 Argon1525
- #2Claude Opus 5.51507
- #11Qwen3.8 Max (0902)1481
- #12GLM-5.31480
- #12MiMo-V2.6-Pro1480
lmarena.ai · 2026-09-25
#6 of 252075 Elo
eqbench.com · 2026-09-24
Hard reasoning
Humanity's Last ExamVerified
#31 of 19842.3%
- #1Claude Fable 5.274.3%
- #2Claude Opus 5.561.4%
- #30Qwen3.8 2.4T A95B42.4%
- #31GLM-5.342.3%
- #32Claude Sonnet 541.3%
artificialanalysis.ai · 2026-09-24
AIME (latest)Estimated
~#2 of 34 (estimated)~99.3%
No published score — estimated from related benchmarks
Agents
τ²-bench (Telecom)Estimated
~#2 of 120 (estimated)~98.3%
No published score — estimated from related benchmarks
OSWorld-VerifiedEstimated
~#3 of 15 (estimated)~83.6%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#2 of 5 (estimated)~58.3%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#4 of 22 (estimated)~90.9%
- #1Atria Dawn Preview92.5%
- #2GPT-6 Astra91.5%
- #3Kimi K391.2%
- ~#4GLM-5.3Estimated90.9%
- #4Claude Opus 590.8%
No published score — estimated from related benchmarks
GDPval-AA (Elo)Verified
#11 of 911653 Elo
- #1Claude Opus 5.51866
- #2Muse Wonder Full1857
- #10Qwen3.8 Max (0902)1668
- #11GLM-5.31653
- #12GLM-5.3 Flash1644
artificialanalysis.ai · 2026-10-07