Free open-weight model APIs
Open-weight models you can call for free via API and, when needed, run on your own hardware without vendor lock-in.
125 models11 providersup to 10.5M contextbest: GLM 5.3Updated:
Read more
Open weights means the parameters are published, but licenses differ: Llama and Gemma have their own terms, while DeepSeek, Qwen, and gpt-oss are more permissive. Check the license before commercial use.
| # | Model ↕ | AA Index ↕ | GPQA ↕ | Context ↕ | Size ↕ | Free at ↕ | |
|---|---|---|---|---|---|---|---|
| 1 |  |
44.8 | 92% | 1M | Frontier |   Together AI, NVIDIA Build |
|
| 2 |  |
43.6 | 94% | 1M | Frontier |  NVIDIA Build, UnoRouter |
|
| 3 | |
41.8 | 91% | 1M | Mid-size |  OrcaRouter, NVIDIA Build, UnoRouter |
|
| 4 |  |
39.5 | — | 1M | Frontier | NVIDIA Build, UnoRouter |
|
| 5 |  |
33.7 | 91% | 262K | Mid-size |    OpenRouter, Groq, Cloudflare Workers AI, UnoRouter |
|
| 6 | |
36.0 | 93% | 1M | Frontier | UnoRouter |
|
| 7 | |
33.7 | 90% | 1M | Frontier | Together AI |
|
| 8 | |
34.3 | 91% | — | Frontier | OrcaRouter |
|
| 9 | 27.8 | 90% | 1M | Small | OpenRouter |
||
| 10 | |
27.0 | 88% | 1M | Frontier | UnoRouter |
|
| 11 |  |
22.9 | 87% | 1M | Frontier | OpenRouter, NVIDIA Build, UnoRouter |
|
| 12 | 25.0 | 87% | 1M | Small | OpenRouter |
||
| 13 | |
27.0 | 91% | — | Frontier | NVIDIA Build |
|
| 14 | 25.3 | 90% | — | Small | OrcaRouter |
||
| 15 | 22.6 | — | 262K | Small |  Nous Portal |
||
| 16 |  |
19.0 | 86% | 262K | Mid-size | OpenRouter, Together AI, NVIDIA Build |
|
| 17 | |
16.7 | 79% | 262K | Mid-size | OpenRouter, Together AI, Cloudflare Workers AI, UnoRouter |
|
| 18 | |
22.2 | 86% | 203K | Frontier | Together AI |
|
| 19 | 19.1 | 78% | 1M | Frontier | Nous Portal |
||
| 20 |  |
19.5 | 81% | 262K | Small | Nous Portal |
|
| 21 | |
19.3 | 85% | 262K | Mid-size | Together AI |
|
| 22 | |
17.7 | 83% | 262K | Frontier | Together AI, UnoRouter |
|
| 23 | |
18.2 | 84% | 262K | Mid-size | Together AI |
|
| 24 |  |
18.6 | 78% | 197K | Frontier | Together AI |
|
| 25 | |
12.9 | 74% | 1M | Mid-size | OpenRouter, NVIDIA Build, UnoRouter |
|
| 26 | |
12.8 | 80% | 262K | Frontier | OpenRouter, NVIDIA Build, UnoRouter |
|
| 27 | |
14.9 | 58% | 131K | Mid-size | Cloudflare Workers AI, UnoRouter |
|
| 28 |  |
— | — | 1M | Frontier | Together AI |
|
| 29 | — | — | 1M | Frontier | Nous Portal |
||
| 30 | |
10.3 | 47% | 262K | Mid-size |  +1OpenRouter, Together AI, NVIDIA Build, TokenRouter, UnoRouter |
|
| 31 | |
14.2 | 75% | 262K | Mid-size | Together AI |
|
| 32 | |
— | — | 256K | Frontier | Cloudflare Workers AI |
|
| 33 |  |
— | — | 256K | Frontier | UnoRouter |
|
| 34 | |
13.7 | 81% | 262K | Small | Together AI |
|
| 35 |  |
11.6 | 78% | 131K | Frontier | Groq, Cloudflare Workers AI, UnoRouter |
|
| 36 | |
10.0 | 61% | 131K | Mid-size | Groq, Cloudflare Workers AI, NVIDIA Build, UnoRouter |
|
| 37 | |
13.4 | 77% | 262K | Frontier | Together AI |
|
| 38 | |
13.1 | 77% | 262K | Small | UnoRouter |
|
| 39 | |
— | — | 131K | Frontier | Together AI |
|
| 40 | |
— | — | 131K | Frontier | Together AI |
|
| 41 | |
— | — | 203K | Frontier | Together AI |
|
| 42 | |
— | — | 203K | Frontier | Together AI |
|
| 43 | |
— | — | 8K | Frontier | Together AI |
|
| 44 | |
— | — | 131K | Frontier | Together AI |
|
| 45 | |
— | — | 24K | Frontier | Cloudflare Workers AI |
|
| 46 | |
12.8 | 80% | 262K | Frontier | Together AI |
|
| 47 | |
11.7 | 70% | 1M | Frontier | Together AI |
|
| 48 | |
12.7 | 79% | 262K | Frontier | Together AI |
|
| 49 | |
11.2 | 76% | 131K | Frontier | Together AI, UnoRouter |
|
| 50 |  |
9.9 | 76% | 262K | Small | OpenRouter, UnoRouter |
|
| 51 | |
9.8 | 71% | 262K | Mid-size | Together AI, Cloudflare Workers AI |
|
| 52 | |
— | — | 131K | Mid-size | Mistral AI, Together AI, NVIDIA Build |
|
| 53 | |
10.0 | 68% | 1M | Frontier | Together AI |
|
| 54 | |
8.1 | 59% | 10.5M | Frontier | Together AI, Cloudflare Workers AI |
|
| 55 | |
9.5 | 59% | 131K | Mid-size | Cloudflare Workers AI, UnoRouter |
|
| 56 | |
— | — | 1M | Mid-size | UnoRouter |
|
| 57 | |
9.6 | 52% | 262K | Frontier | Together AI |
|
| 58 | — | — | 262K | Small |  OpenRouter, Vercel AI Gateway, Nous Portal, UnoRouter |
||
| 59 | |
9.5 | 67% | 262K | Mid-size | UnoRouter |
|
| 60 | |
10.0 | 67% | 131K | Frontier | UnoRouter |
|
| 61 | |
— | — | 262K | Mid-size | UnoRouter |
|
| 62 | 8.4 | 56% | 128K | Small | OpenRouter, UnoRouter |
||
| 63 | |
5.4 | 22% | 131K | Mid-size | Together AI, Cloudflare Workers AI, NVIDIA Build, UnoRouter |
|
| 64 | |
8.9 | 76% | 262K | Mid-size | Together AI |
|
| 65 | |
8.4 | 62% | 80K | Mid-size | Cloudflare Workers AI, UnoRouter |
|
| 66 | |
8.8 | 67% | 262K | Small | Together AI |
|
| 67 | |
— | — | 16K | Mid-size | Together AI |
|
| 68 | |
— | — | 131K | Mid-size | Together AI |
|
| 69 | |
— | — | 131K | Mid-size | Together AI |
|
| 70 | |
— | — | 131K | Mid-size | Together AI |
|
| 71 | |
— | — | 131K | Mid-size | Together AI |
|
| 72 | |
— | — | 8K | Mid-size | Together AI |
|
| 73 | |
— | — | 128K | Mid-size | Cloudflare Workers AI |
|
| 74 | |
— | — | 128K | Mid-size | Cloudflare Workers AI |
|
| 75 | |
— | — | 128K | Mid-size | UnoRouter |
|
| 76 | |
— | — | 33K | Mid-size | UnoRouter |
|
| 77 | |
— | — | 25K | Mid-size | UnoRouter |
|
| 78 | |
8.9 | 64% | 16K | Mid-size | Together AI |
|
| 79 | |
8.9 | 58% | 131K | Small | Together AI |
|
| 80 | |
8.9 | 76% | — | Mid-size | NVIDIA Build |
|
| 81 | |
8.7 | 43% | 131K | Mid-size | Together AI |
|
| 82 | |
8.6 | 66% | 41K | Mid-size | Together AI |
|
| 83 | |
8.6 | 67% | 41K | Mid-size | Together AI |
|
| 84 | |
6.9 | 26% | 32K | Small | Together AI, Cloudflare Workers AI |
|
| 85 | |
5.8 | 38% | 256K | Mid-size | Mistral AI, UnoRouter |
|
| 86 | |
8.2 | 60% | 2K | Mid-size | Together AI |
|
| 87 | |
6.7 | 42% | 131K | Mid-size | Cloudflare Workers AI, UnoRouter |
|
| 88 | |
6.4 | 43% | 16K | Frontier | Together AI, NVIDIA Build |
|
| 89 | |
6.9 | 46% | 262K | Small | Together AI |
|
| 90 | |
7.8 | 43% | 131K | Small | Together AI |
|
| 91 | |
7.7 | 50% | 131K | Frontier | Together AI |
|
| 92 | |
7.7 | 49% | 131K | Frontier | Together AI |
|
| 93 | |
7.6 | 68% | 66K | Frontier | Together AI |
|
| 94 | |
7.5 | 73% | — | Frontier | NVIDIA Build |
|
| 95 | |
7.5 | 57% | 131K | Mid-size | UnoRouter |
|
| 96 | |
7.4 | 57% | 131K | Small | UnoRouter |
|
| 97 | |
7.3 | 52% | 131K | Frontier | Together AI |
|
| 98 | |
7.3 | 59% | 41K | Small | Together AI |
|
| 99 | |
6.9 | 47% | 131K | Mid-size | Together AI |
|
| 100 | |
6.6 | 41% | 16K | Frontier | Together AI |
|
| 101 | |
4.8 | 29% | 66K | Small | Together AI, NVIDIA Build |
|
| 102 | |
4.8 | 20% | 131K | Small | Together AI, Cloudflare Workers AI |
|
| 103 | — | — | 262K | Small | Vercel AI Gateway |
||
| 104 | |
5.7 | 26% | 80K | Small | Cloudflare Workers AI |
|
| 105 | 5.6 | 24% | 4K | Small | Together AI |
||
| 106 |
Sarvam M
Open weights
|
5.3 | 42% | 33K | Small | Together AI |
|
| 107 | |
5.2 | 36% | 41K | Small | Together AI |
|
| 108 | |
5.1 | 22% | 33K | Small | Together AI |
|
| 109 | 5.1 | 34% | 131K | Small | UnoRouter |
||
| 110 | |
— | — | 131K | Small | Together AI |
|
| 111 | |
— | — | 131K | Small | Together AI |
|
| 112 | |
— | — | 33K | Small | Together AI |
|
| 113 | |
— | — | 131K | Small | Together AI |
|
| 114 | |
— | — | 131K | Small | Together AI |
|
| 115 | |
— | — | 131K | Small | Together AI |
|
| 116 | — | — | 131K | Small | Cloudflare Workers AI |
||
| 117 | |
— | — | 4K | Small | Cloudflare Workers AI |
|
| 118 | |
— | — | 8K | Small | NVIDIA Build |
|
| 119 | |
— | — | 131K | Small | UnoRouter |
|
| 120 | |
4.9 | 43% | 66K | Mid-size | Together AI |
|
| 121 | |
4.8 | 24% | 33K | Small | Together AI |
|
| 122 | |
4.8 | 23% | 41K | Small | Together AI |
|
| 123 | |
3.8 | 35% | — | Mid-size | NVIDIA Build |
|
| 124 | |
— | — | 128K | Mid-size | OpenRouter |
|
| 125 | |
— | — | 131K | Mid-size | UnoRouter |
No models match your filters.
Benchmarks: Artificial Analysis (September 30, 2026). A dash means the model is not on their leaderboard; we never estimate scores ourselves.
FAQ
Which open-weight models are free right now?
125 models are free right now. Leaders by the Artificial Analysis index: GLM 5.3, Kimi K3, GLM 5.3 Flash, DeepSeek V4.1 Flash, Qwen3.8 27B, DeepSeek V4 Pro.
Where can I use GLM 5.3 for free?
GLM 5.3 is free at Together AI, NVIDIA Build. Each provider is OpenAI-compatible — just switch the base URL and key.
What are the limits of free access?
Open weights means the parameters are published, but licenses differ: Llama and Gemma have their own terms, while DeepSeek, Qwen, and gpt-oss are more permissive. Check the license before commercial use.