GLM-5.3-Flash
Z.ai · released
A much cheaper, natively multimodal sibling of GLM-5.3 on a new, smaller architecture. It accepts text, images, video and files, keeps the 1M-token context and targets everyday coding agents and office-document work.
Best for: Cheap multimodal agents · MIT-licensed self-hosting · Everyday coding
- Intelligence Index
- 41.8
- #16 of 38
- Price per 1M tokens
- $0.15 in $0.50 out
- Blended $0.24 at 3:1
- Context window
- 1M
- Up to 128K output
- Output speed
- 51 tokens/s
- $0.25 measured cost per task
Capability profile
One 0–100 score per area, placing this model between the weakest and strongest tracked model across that area's benchmarks.
- Overall68 · #15 of 38
Based on 2 of 3 benchmarks
- Coding55 · #17 of 39
Based on 3 of 5 benchmarks
- Agents & knowledge work71 · #10 of 40
Based on 2 of 6 benchmarks
- Reasoning & science53 · #20 of 39
Based on 2 of 2 benchmarks
- Knowledge & reliability67 · #17 of 39
Based on 2 of 2 benchmarks
- Vision77 · #9 of 31
Based on 1 of 2 benchmarks
Benchmark results
All third-party results on record. Rank is among the tracked models that have been tested on that benchmark.
Overall
#16 of 38· Artificial Analysis
- Vals IndexNot tested
- Arena · Text1473 ±5
#14 of 33· 22,971 votes· Arena
Coding
- Terminal-Bench 4.032.8%
#13 of 38· Artificial Analysis
- SciCode51.6%
#21 of 38· Artificial Analysis
- Arena · WebDev1610 ±7
#17 of 34· 12,868 votes· Arena
- Vibe Code BenchNot tested
- Code MigrationNot tested
Agents & knowledge work
- GDPval-AA57.3
#9 of 39· Artificial Analysis
- Arena · Agent-0.4 ±0.5
#20 of 29· Arena
- Finance Agent v2Not tested
- Excel Modeling BenchmarkNot tested
- Legal Research BenchNot tested
- Tax Agent BenchNot tested
Reasoning & science
- Humanity's Last Exam39.9%
#19 of 39· Artificial Analysis
- CritPt15.4%
#20 of 39· Artificial Analysis
- ARC-AGI-2saturating65.8%
#9 of 16· Max effort· $0.09 per task· ARC Prize
- ARC-AGI-3Not tested
- GPQA Diamondsaturating91.2%
#13 of 27· Artificial Analysis
Knowledge & reliability
#17 of 39· Artificial Analysis
- AA-LCR80.0%
#22 of 39· Artificial Analysis
Vision
- MMMU-ProNot tested
- Arena · Vision1278 ±9
#8 of 19· 5,479 votes· Arena
Reported by Z.ai
Numbers the lab published at launch, run on its own harness and settings. Useful context, but not independently verified and never used in the rankings on this site.
- OSWorld 2.059.1%
Specifications
- Lab
- Z.ai
- Released
- Aug 26, 2026
- Availability
- Generally available
- Context window
- 1,000,000 tokens
- Max output
- 128,000 tokens
- Accepts
- Text, Images, Video, PDFs
- Knowledge cutoff
- Not published
- Weights
- Open
- Licence
- MIT
- Parameters
- 320B total · 18B active
- Reasoning control
- Always on · effort low to max
- API model ID
glm-5.3-flash
Pricing
- Input / 1M tokens
- $0.15
- Cached input / 1M
- $0.03
- Output / 1M tokens
- $0.50
- Blended (3:1)
- $0.24
- Measured cost / task
- $0.25
Official sources
Closest alternatives
The models nearest to this one on overall score.
23 benchmarks tracked. See methodology and sources.