Kostenlose Modelle mit langem Kontext
Modelle mit einem Kontextfenster von 256K Token oder mehr — für RAG ohne Chunking, Reviews ganzer Repositories und lange Dokumente.
84 Modelle13 Anbieterbis zu 10.5M Kontextbeste: GLM 5.3 FlashAktualisiert:
Weiterlesen
Kostenlose Routen begrenzen den Kontext oft unter das Maximum des Modells. Die Spalte zeigt das anbieterspezifische Limit.
| # | Modell ↕ | AA Index ↕ | GPQA ↕ | Kontext ↕ | Größe ↕ | Kostenlos bei ↕ | |
|---|---|---|---|---|---|---|---|
| 1 |  |
8.1 | 59% | 10.5M | Sehr groß |    Together AI, Cloudflare Workers AI, UnoRouter |
|
| 2 |
Mindshub Air
New
|
— | — | 1.1M | Klein | UnoRouter |
|
| 3 |  |
44.8 | 92% | 1M | Sehr groß | Together AI |
|
| 4 |  |
40.9 | 95% | 1M | Mittelgroß | Google AI Studio |
|
| 5 | |
33.7 | 90% | 1M | Sehr groß | Together AI, UnoRouter |
|
| 6 |  |
22.9 | 87% | 1M | Sehr groß |   OpenRouter, NVIDIA Build, UnoRouter |
|
| 7 | |
— | — | 1M | Sehr groß | Together AI |
|
| 8 | |
— | — | 1M | Mittelgroß |  OpenRouter, UnoRouter, OpenCode Zen |
|
| 9 |  |
11.7 | 70% | 1M | Sehr groß | Together AI |
|
| 10 | |
10.0 | 68% | 1M | Sehr groß | Together AI |
|
| 11 | — | — | 1M | Klein | OpenRouter, OpenCode Zen |
||
| 12 | — | — | 1M | Klein |  Vercel AI Gateway, OpenCode Zen |
||
| 13 | |
41.8 | 91% | 1M | Mittelgroß |  OrcaRouter, NVIDIA Build, UnoRouter, OpenCode Zen |
|
| 14 |  |
43.7 | — | 1M | Sehr groß | NVIDIA Build, UnoRouter |
|
| 15 | |
41.8 | 91% | 1M | Mittelgroß | UnoRouter |
|
| 16 | |
34.0 | 93% | 1M | Mittelgroß | UnoRouter |
|
| 17 | |
26.3 | 90% | 1M | Mittelgroß | UnoRouter |
|
| 18 | |
22.2 | 84% | 1M | Klein | UnoRouter |
|
| 19 | |
12.9 | 74% | 1M | Mittelgroß | +1OpenRouter, NVIDIA Build, Vercel AI Gateway, UnoRouter, OpenCode Zen |
|
| 20 | |
— | — | 1M | Sehr groß | NVIDIA Build, UnoRouter, OpenCode Zen |
|
| 21 | |
15.6 | 82% | 1M | Klein | UnoRouter |
|
| 22 |  |
— | — | 1M | Mittelgroß | UnoRouter |
|
| 23 | |
— | — | 1M | Mittelgroß | UnoRouter |
|
| 24 | |
— | — | 1M | Mittelgroß | UnoRouter |
|
| 25 |  |
— | — | 1M | Mittelgroß | UnoRouter |
|
| 26 | |
8.5 | 63% | 1M | Klein | UnoRouter |
|
| 27 |  |
28.2 | 89% | 524K | Sehr groß | UnoRouter |
|
| 28 |
Solar Mini4
Upstage · New
|
— | — | 524K | Klein | UnoRouter |
|
| 29 | — | — | 524K | Klein | UnoRouter |
||
| 30 |
Dots3-Note Preview
Dots Studio
|
— | — | 512K | Klein | OpenRouter, UnoRouter |
|
| 31 | 
Command A Plus 05 2026
Cohere · 111B · Legacy
|
— | — | 436K | Sehr groß | Cohere |
|
| 32 | |
— | — | 289K | Sehr groß | Cohere |
|
| 33 |
Command A 03 2025
Cohere · 111B · Legacy
|
— | — | 288K | Sehr groß | Cohere |
|
| 34 | |
33.7 | 91% | 262K | Mittelgroß |  Groq, Cloudflare Workers AI, UnoRouter |
|
| 35 |  |
19.0 | 86% | 262K | Mittelgroß | OpenRouter, Together AI, NVIDIA Build, UnoRouter |
|
| 36 | |
16.7 | 79% | 262K | Mittelgroß | OpenRouter, Together AI, Cloudflare Workers AI, UnoRouter |
|
| 37 | |
19.3 | 85% | 262K | Mittelgroß | Together AI |
|
| 38 | |
18.2 | 84% | 262K | Mittelgroß | Together AI |
|
| 39 | |
17.7 | 83% | 262K | Sehr groß | Together AI |
|
| 40 | |
14.2 | 75% | 262K | Mittelgroß | Together AI, UnoRouter |
|
| 41 | |
13.7 | 81% | 262K | Klein | Together AI, UnoRouter |
|
| 42 | |
12.8 | 80% | 262K | Sehr groß | NVIDIA Build, UnoRouter |
|
| 43 | — | — | 262K | Sehr groß | UnoRouter |
||
| 44 | — | — | 262K | Sehr groß | UnoRouter |
||
| 45 | |
— | — | 262K | Sehr groß | UnoRouter |
|
| 46 | |
13.4 | 77% | 262K | Sehr groß | Together AI |
|
| 47 | |
13.1 | 77% | 262K | Klein | UnoRouter |
|
| 48 | |
12.8 | 80% | 262K | Sehr groß | Together AI |
|
| 49 | |
12.7 | 79% | 262K | Sehr groß | Together AI |
|
| 50 | |
9.8 | 71% | 262K | Mittelgroß | Together AI, Cloudflare Workers AI |
|
| 51 |
Xing4.0 29B A4B
XingChen · 29B · 4B
|
— | — | 262K | Mittelgroß | OpenRouter, UnoRouter |
|
| 52 | |
9.6 | 52% | 262K | Sehr groß | Together AI |
|
| 53 | |
9.5 | 67% | 262K | Mittelgroß | UnoRouter |
|
| 54 | — | — | 262K | Mittelgroß | Together AI |
||
| 55 |
Holo3 35B A3B
35B · 3B
|
— | — | 262K | Mittelgroß | Together AI |
|
| 56 | — | — | 262K | Mittelgroß | UnoRouter |
||
| 57 | — | — | 262K | Mittelgroß | UnoRouter |
||
| 58 | |
— | — | 262K | Mittelgroß | UnoRouter |
|
| 59 | |
8.9 | 76% | 262K | Mittelgroß | Together AI |
|
| 60 | |
8.8 | 67% | 262K | Klein | Together AI |
|
| 61 | |
6.9 | 46% | 262K | Klein | Together AI |
|
| 62 |
Flint
Recursive
|
— | — | 262K | Klein | OpenRouter, UnoRouter |
|
| 63 | — | — | 262K | Klein | NVIDIA Build, UnoRouter |
||
| 64 |
LFM3-1.7B
Liquid AI · 1.7B · New
|
— | — | 262K | Klein | OpenRouter |
|
| 65 | — | — | 262K | Klein | Cloudflare Workers AI |
||
| 66 | — | — | 262K | Klein | UnoRouter |
||
| 67 | — | — | 262K | Klein | UnoRouter |
||
| 68 | — | — | 262K | Klein | UnoRouter |
||
| 69 | — | — | 262K | Klein | Cloudflare Workers AI |
||
| 70 | |
21.4 | 86% | 262K | Sehr groß | UnoRouter |
|
| 71 | |
21.4 | 84% | 262K | Mittelgroß | UnoRouter |
|
| 72 | |
— | — | 262K | Mittelgroß | UnoRouter |
|
| 73 | |
— | — | 262K | Mittelgroß | UnoRouter |
|
| 74 | |
— | — | 262K | Mittelgroß | UnoRouter |
|
| 75 | |
— | — | 262K | Mittelgroß | UnoRouter |
|
| 76 | |
— | — | 262K | Mittelgroß | UnoRouter |
|
| 77 |  |
5.8 | 38% | 262K | Mittelgroß | Mistral AI, UnoRouter |
|
| 78 | — | — | 262K | Klein | UnoRouter |
||
| 79 | |
10.3 | 47% | 256K | Mittelgroß |  +1OpenRouter, Together AI, NVIDIA Build, TokenRouter, UnoRouter |
|
| 80 | |
— | — | 256K | Sehr groß | Cloudflare Workers AI |
|
| 81 | |
9.9 | 76% | 256K | Klein | OpenRouter, UnoRouter |
|
| 82 |
Cohere Command A
Cohere · 111B · Legacy
|
— | — | 256K | Sehr groß | UnoRouter |
|
| 83 | |
— | — | 256K | Klein | Mistral AI |
|
| 84 | |
— | — | 256K | Klein | UnoRouter |
Keine Modelle entsprechen Ihren Filtern.
Benchmarks: Artificial Analysis (24. September 2026). Ein Strich bedeutet, dass das Modell nicht in deren Rangliste steht; wir schätzen Werte nie selbst.
FAQ
Welches ist das beste kostenlose Modell für langer kontext?
GLM 5.3 Flash führt derzeit — kostenlos bei OrcaRouter, NVIDIA Build, UnoRouter, OpenCode Zen. Danach folgen Step 5 Preview, GLM 5.3, GLM 5.3 Flash Thinking.