Gemini 3 Flash PreviewvsGLM 5.2
Gemini 3 Flash Preview is cheaper than GLM 5.2 at $0.5/M vs $1.4/M input tokens.
| Spec | Gemini 3 Flash Preview | GLM 5.2 |
|---|---|---|
| Input price | $0.5/M tokens | $1.4/M tokens |
| Output price | $3/M tokens | $4.4/M tokens |
| Context window | 1.0M tokens | 1.0M tokens |
| Weights | Closed | Open |
| Free API (OpenRouter) | No | No |
| Released | Dec 2025 | Jun 2026 |
Our Verdict
GLM 5.2
Gemini 3 Flash PreviewRunner-up
Not enough votes to call it. On the specs, GLM 5.2 has the edge: bigger model tier, newer.
Too close to call
API pricing
Cost per 1M tokens
Gemini 3 Flash Preview
Input
$0.50
2.8× cheaper
Output
$3.00
1.5× cheaper
GLM 5.2
Input
$1.40
Output
$4.40
Gemini 3 Flash Preview is cheaper on both: 2.8× input, 1.5× output.
Where to run it
26 hosts, cheapest first
Gemini 3 Flash Preview2 hosts
GLM 5.224 hosts
DeepInfrafp4$0.49 in$1.56 out1M99.4% upAmbientfp8$0.60 in$2.00 out203k97.9% upStreamLakefp8$0.68 in$2.13 out1M99.7% upNovitafp8$0.68 in$2.15 out1M100% upDigitalOcean$0.70 in$2.20 out262k99.6% upCoreWeavefp4$0.76 in$2.42 out1M99.9% up
18 more hostsFewer hosts
AtlasCloudfp8$0.94 in$2.95 out1M99.9% upAlibaba Cloudfp8$0.97 in$3.04 out1M99.9% upSiliconFlowfp8$1.19 in$3.74 out1M99.8% upInceptronfp4$1.25 in$2.99 out1M99.3% upPhalafp8$1.26 in$3.00 out1M99.8% upBaidu Qianfanfp8$1.40 in$4.40 out1M99.8% upBasetenfp8$1.40 in$4.40 out1M99.7% upCloudflare Workers AI$1.40 in$4.40 out262k100% upCrusoefp8$1.40 in$4.40 out1M99.9% upFireworks$1.40 in$4.40 out1M99.9% upFriendli$1.40 in$4.40 out1M99.9% upGMI Cloudfp8$1.40 in$4.40 out1M99.5% upMistral$1.40 in$4.40 out1M99.9% upParasailfp4$1.40 in$4.40 out262k99.9% upTogether$1.40 in$4.40 out512k98.3% upVenicefp8$1.40 in$4.40 out1M99.6% upZ.aifp8$1.40 in$4.40 out1M99.7% upDecartfp4$1.89 in$5.94 out1M99.8% up
Per million tokens. Prices and uptime via OpenRouter, checked 8 Sep 2026.
Writing DNA
Style Comparison
Similarity
62%
GLM 5.2 uses 2.8x more transitions
Gemini 3 Flash Preview
GLM 5.2
57%Vocabulary56%
18wSentence Length18w
0.41Hedging0.27
7.4Bold3.7
4.2Lists2.8
0.00Emoji0.00
0.77Headings0.56
0.08Transitions0.24
Based on 28 + 27 text responses
Research
What we learned reading every model
FAQ
Common questions
Keep exploring
