vs
2:13
overall capability wins (0 ties across 15 scored). GLM-5.3 Flash leads overall.
- Reasoning2:3
- Coding0:4
- Arena0:1
- Agents0:1
- Tool use0:2
- Composite indices0:2
- Pricing0:3
- Speed2:0
- Specs0:2
Verdict
GLM-5.3 Flash leads coding (widest gap: +19.0 on DeepSWE); GLM-5.3 Flash ranks higher on LMArena (+3 Elo); GLM-5.3 Flash offers ~5.0x larger context (1M vs 200K).
- GLM-5.3 Flash leads coding (widest gap: +19.0 on DeepSWE)
- GLM-5.3 Flash ranks higher on LMArena (+3 Elo)
- GLM-5.3 Flash offers ~5.0x larger context (1M vs 200K)
- GLM-5.3 Flash allows ~3.9x more max output (128K vs 33K)
- GLM-5.3 Flash is ~9.1x cheaper on a blended token basis than GLM-5.2
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- GLM-5.2
- $2.50
- GLM-5.3 Flash
- $0.28
GLM-5.3 Flash is about 9.1x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Zhipu AI
Zhipu AI
Family
GLM
GLM
License
MIT
MIT
Open weights
Yes
Yes
Release
Jun 1, 2026
Aug 26, 2026
Knowledge cutoff
-
-
API / provider
Zhipu / Z.ai
Zhipu / Z.ai
Modalities
text, image → text
text, image, video → text
Specs
Context window
200K
1M
Max output
33K
128K
Parameters
—
320B (18B act.)
Pricing
Input $/1M
$1.40
$0.15
Output $/1M
$4.40
$0.50
Blended $/1M (3∶1)
$2.15
$0.24
Speed
tok/s
168
49.8
TTFT (s)
0.35
1.51
Reasoning
MMLU-Pro
—
—
GPQA Diamond
89.5
91.2
Humanity's Last Exam
41.1
39.9
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
24.3
27.5
AA-LCR v1.1
78.3
80.0
CritPt
20.9
15.4
MMMU-Pro
—
—
IFBench
73.3
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
77.9
84.3
Aider Polyglot
—
—
Terminal-Bench 3
4.6
—
BigCodeBench
—
—
SciCode
51.2
51.6
CursorBench
55.0
—
SWE-Rebench
57.0
—
NL2Repo-Bench
—
—
DeepSWE
44.0
63.0
WebDev Arena
1,592
1,607
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,472
1,475
| Metric | GLM-5.2 | GLM-5.3 Flash | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Zhipu AI | Zhipu AI | — |
| Family | GLM | GLM | — |
| License | MIT | MIT | — |
| Open weights | Yes | Yes | — |
| Release | Jun 1, 2026 | Aug 26, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Zhipu / Z.ai | Zhipu / Z.ai | — |
| Modalities | text, image → text | text, image, video → text | — |
| Specs | |||
| Context window | 200K | 1M | B +800K |
| Max output | 33K | 128K | B +95K |
| Parameters | — | 320B (18B act.) | — |
| Pricing | |||
| Input $/1M | $1.40 | $0.15 | B +$1.25 |
| Output $/1M | $4.40 | $0.50 | B +$3.90 |
| Blended $/1M (3∶1) | $2.15 | $0.24 | B +$1.91 |
| Speed | |||
| tok/s | 168 | 49.8 | A +118 |
| TTFT (s) | 0.35 | 1.51 | A +1.16 |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 89.5 | 91.2 | B +1.7 pts |
| Humanity's Last Exam | 41.1 | 39.9 | A +1.2 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 24.3 | 27.5 | B +3.2 pts |
| AA-LCR v1.1 | 78.3 | 80.0 | B +1.7 pts |
| CritPt | 20.9 | 15.4 | A +5.5 pts |
| MMMU-Pro | — | — | — |
| IFBench | 73.3 | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 77.9 | 84.3 | B +6.4 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | 4.6 | — | — |
| BigCodeBench | — | — | — |
| SciCode | 51.2 | 51.6 | B +0.4 pts |
| CursorBench | 55.0 | — | — |
| SWE-Rebench | 57.0 | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | 44.0 | 63.0 | B +19.0 pts |
| WebDev Arena | 1,592 | 1,607 | B +15 Elo |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,472 | 1,475 | B +3 Elo |
GLM-5.2
- Context
- 200K / 33K out
- Parameters
- —
- Price
- $1.40 / $4.40
- Speed
- 168 tok/s · 0.35s TTFT
- Modalities
- text, image → text
- License
- MIT
Open MIT GLM-5.2 - competitive open Arena Elo and strong Chinese/English bilingual performance.
GLM-5.3 Flash
- Context
- 1M / 128K out
- Parameters
- 320B (18B act.)
- Price
- $0.15 / $0.50
- Speed
- 49.8 tok/s · 1.51s TTFT
- Modalities
- text, image, video → text
- License
- MIT
First natively multimodal GLM-5 (text/image/video in), MIT open weights, 320B total / 18B active MoE with 1M context, released Aug 26, 2026. List pricing $0.15/$0.50 per 1M (cache-hit ~$0.03), halved by a launch promo through Sept 9, 2026. Benchmark figures are Z.ai's self-reported launch numbers (DeepSWE 63.4, AutomationBench 48.8, Terminal-Bench 2.1 84.3); Artificial Analysis places it at 57 on its Intelligence Index.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.