Free models and APIs on Together AI in 2026

Together AI offers a unified API for open models, including selected models with zero-cost access.

82 modelsup to 10.5M contextbest: GLM 5.3Updated: October 24, 2026 at 06:40 PM UTC

Together AI

Online at the last check

Only selected models are priced at zero; there is no general free trial.

Base URL: https://api.together.xyz/v1
Compatibility: OpenAI Chat Completions
Card: Check with the provider
Limits: Fair use (depends on server load)

#	Model ↕	AA Index ↕	Context ↕	Size ↕	Limit
1	GLM 5.3 `zai-org/GLM-5.3-FP8`	44.8	1M	Frontier	Fair use (depends on server load)	Open
2	GLM 5.3 `zai-org/GLM-5.3-FP8-Lora`	44.8	1M	Frontier	Fair use (depends on server load)	Open
3	Gemma 4 26B A4B `google/gemma-4-26B-A4B-it`	16.7	262K	Mid-size	Fair use (depends on server load)	Open
4	Qwen 3.6 35B `Qwen/Qwen3.6-35B-A3B-FP8`	18.2	262K	Mid-size	Fair use (depends on server load)	Open
5	Qwen 3.6 35B `Qwen/Qwen3.6-35B-A3B-Lora`	18.2	262K	Mid-size	Fair use (depends on server load)	Open
6	Qwen 3.5 122B `Qwen/Qwen3.5-122B-A10B-FP8`	17.7	262K	Frontier	Fair use (depends on server load)	Open
7	Gemma 4 31B `google/gemma-4-31B-it-lora`	14.7	262K	Mid-size	Fair use (depends on server load)	Open
8	MiniMax M1 40K `MiniMaxAI/MiniMax-M1-40k`	—	1M	Frontier	Fair use (depends on server load)	Open
9	MiniMax M1 80K `MiniMaxAI/MiniMax-M1-80k`	—	1M	Frontier	Fair use (depends on server load)	Open
10	Llama 4 Maverick Instruct Fp4 `meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP4`	—	1M	Frontier	Fair use (depends on server load)	Open
11	GLM 5.2 `zai-org/GLM-5.2-FP8`	—	1M	Frontier	Fair use (depends on server load)	Open
12	GLM 5.2 `zai-org/GLM-5.2-FP8-Lora`	—	1M	Frontier	Fair use (depends on server load)	Open
13	Nemotron 3 Nano Omni `nvidia/nemotron-3-nano-omni-30b-a3b-reasoning-fp8`	10.3	131K	Mid-size	Fair use (depends on server load)	Open
14	Gemma 4 12B `google/gemma-4-12B-it`	14.2	262K	Mid-size	Fair use (depends on server load)	Open
15	Qwen3 235B `Qwen/Qwen3-235B-A22B-Instruct-2507-FP8`	—	262K	Frontier	Fair use (depends on server load)	Open
16	Qwen3 Coder 30B `Qwen/Qwen3-Coder-30B-A3B-Instruct`	—	262K	Frontier	Fair use (depends on server load)	Open
17	Qwen3 VL 235B `Qwen/Qwen3-VL-235B-A22B-Instruct-FP8`	—	262K	Frontier	Fair use (depends on server load)	Open
18	Nvidia Nemotron 3 Super `nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8`	—	262K	Frontier	Fair use (depends on server load)	Open
19	Nvidia Nemotron 3 Super `nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16`	—	262K	Frontier	Fair use (depends on server load)	Open
20	Qwen 3.5 9B `Qwen/Qwen3.5-9B-FP8`	13.3	262K	Small	Fair use (depends on server load)	Open
21	Llama 3.1 70B `nim/meta/llama-3.1-70b-instruct`	—	16K	Frontier	Fair use (depends on server load)	Open
22	Cogito V1 Preview Llama 70B `deepcogito/cogito-v1-preview-llama-70B`	—	131K	Frontier	Fair use (depends on server load)	Open
23	Cogito V1 Preview Llama 70B `deepcogito/cogito-v1-preview-llama-70B-Turbo`	—	131K	Frontier	Fair use (depends on server load)	Open
24	Qwen2.5 72B `Qwen/Qwen2.5-72B`	—	131K	Frontier	Fair use (depends on server load)	Open
25	Meta Llama 3.1 70B `meta-llama/Meta-Llama-3.1-70B`	—	131K	Frontier	Fair use (depends on server load)	Open
26	GLM 4 5v `zai-org/GLM-4.5V`	—	66K	Frontier	Fair use (depends on server load)	Open
27	MiniMax M2 `MiniMaxAI/MiniMax-M2`	—	197K	Frontier	Fair use (depends on server load)	Open
28	GLM 4.7 Fp4 `zai-org/GLM-4.7-fp4`	—	203K	Frontier	Fair use (depends on server load)	Open
29	GLM 5 Fp4 `zai-org/GLM-5-FP4`	—	203K	Frontier	Fair use (depends on server load)	Open
30	MiniMax M2.5 Fp4 `MiniMaxAI/MiniMax-M2.5-FP4`	—	8K	Frontier	Fair use (depends on server load)	Open
31	GLM 4.7 `zai-org/GLM-4.7-FP8`	—	203K	Frontier	Fair use (depends on server load)	Open
32	GLM Ocr `zai-org/GLM-OCR`	—	131K	Frontier	Fair use (depends on server load)	Open
33	Qwen3 Next 80B `Qwen/Qwen3-Next-80B-A3B-Instruct-FP8`	11.2	—	Frontier	Fair use (depends on server load)	Open
34	Mistral Nemo `nim/nv-mistralai/mistral-nemo-12b-instruct`	—	16K	Mid-size	Fair use (depends on server load)	Open
35	Qwen3 30B `Qwen/Qwen3-30B-A3B`	—	41K	Mid-size	Fair use (depends on server load)	Open
36	Llama 4 Scout `meta-llama/Llama-4-Scout-17B-16E`	6.5	262K	Frontier	Fair use (depends on server load)	Open
37	Qwen3 30B `Qwen/Qwen3-30B-A3B-Instruct-2507-Lora`	—	262K	Mid-size	Fair use (depends on server load)	Open
38	Llama 4 Scout `meta-llama/Llama-4-Scout-17B-16E-Instruct-FP8-Lora`	6.5	10.5M	Frontier	Fair use (depends on server load)	Open
39	Ternary Bonsai 27B `Prism-ML/Ternary-Bonsai-27B`	—	262K	Mid-size	Fair use (depends on server load)	Open
40	Qwen 3.5 35B `Qwen/Qwen3.5-35B-A3B`	—	262K	Mid-size	Fair use (depends on server load)	Open
41	Holo3 35B A3B `Hcompany/Holo3-35B-A3B`	—	262K	Mid-size	Fair use (depends on server load)	Open
42	Qwen 3.5 35B `Qwen/Qwen3.5-35B-A3B-Lora`	—	262K	Mid-size	Fair use (depends on server load)	Open
43	Llama 3.2 11B Vision `nim/meta/llama-3.2-11b-vision-instruct`	5.4	16K	Mid-size	Fair use (depends on server load)	Open
44	Nvidia Nemotron 3 Nano `nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16`	8.9	262K	Mid-size	Fair use (depends on server load)	Open
45	Llama 3.3 70B `nim/meta/llama-3.3-70b-instruct`	7.7	16K	Frontier	Fair use (depends on server load)	Open
46	Llama 3.3 70B `meta-llama/Llama-3.3-70B-Instruct`	7.7	131K	Frontier	Fair use (depends on server load)	Open
47	Llama 3.3 70B `meta-llama/Llama-3.3-70B-Instruct-FP8-Lora`	7.7	131K	Frontier	Fair use (depends on server load)	Open
48	Mixtral 8x7b Instruct V01 `nim/mistralai/mixtral-8x7b-instruct-v01`	—	16K	Mid-size	Fair use (depends on server load)	Open
49	Llama 3.3 Nemotron Super 49B `nim/nvidia/llama-3.3-nemotron-super-49b-v1`	—	16K	Mid-size	Fair use (depends on server load)	Open
50	Gemma 3 27B `google/gemma-3-27b-it`	—	66K	Mid-size	Fair use (depends on server load)	Open
51	Cogito V1 Preview Qwen 32B `deepcogito/cogito-v1-preview-qwen-32B`	—	131K	Mid-size	Fair use (depends on server load)	Open
52	Cogito V1 Preview Qwen 14B `deepcogito/cogito-v1-preview-qwen-14B`	—	131K	Mid-size	Fair use (depends on server load)	Open
53	Qwen2.5 32B `Qwen/Qwen2.5-32B`	—	131K	Mid-size	Fair use (depends on server load)	Open
54	Qwen2.5 14B `Qwen/Qwen2.5-14B`	—	131K	Mid-size	Fair use (depends on server load)	Open
55	Qwen2.5 32B `Qwen/Qwen2.5-32B-Instruct`	—	33K	Mid-size	Fair use (depends on server load)	Open
56	Deepcoder 14B Preview `agentica-org/DeepCoder-14B-Preview`	—	131K	Mid-size	Fair use (depends on server load)	Open
57	Devstral Small `mistralai/Devstral-Small-2505`	—	131K	Mid-size	Fair use (depends on server load)	Open
58	Qwen3 14B `Qwen/Qwen3-14B`	—	2K	Mid-size	Fair use (depends on server load)	Open
59	Qwen3 32B `Qwen/Qwen3-32B`	—	41K	Mid-size	Fair use (depends on server load)	Open
60	Medgemma 27B Text `google/medgemma-27b-text-it`	—	131K	Mid-size	Fair use (depends on server load)	Open
61	DeepSeek Ocr 2 `deepseek-ai/DeepSeek-OCR-2`	—	8K	Mid-size	Fair use (depends on server load)	Open
62	Gemma 3 27B `google/gemma-3-27b-it-lora`	—	—	Mid-size	Fair use (depends on server load)	Open
63	Gemma 4 E4b `google/gemma-4-E4B-it`	8.9	131K	Small	Fair use (depends on server load)	Open
64	Magistral Small `mistralai/Magistral-Small-2506`	8.6	41K	Mid-size	Fair use (depends on server load)	Open
65	Llama 3.2 90B Vision `nim/meta/llama-3.2-90b-vision-instruct`	6.4	16K	Frontier	Fair use (depends on server load)	Open
66	Qwen 3.5 2B `Qwen/Qwen3.5-2B-Lora`	6.9	262K	Small	Fair use (depends on server load)	Open
67	Gemma 4 E2b `google/gemma-4-E2B-it`	7.8	131K	Small	Fair use (depends on server load)	Open
68	Llama 3.1 Nemotron 70B `nim/nvidia/llama-3.1-nemotron-70b-instruct`	—	16K	Frontier	Fair use (depends on server load)	Open
69	Mixtral 8x22b `mistralai/Mixtral-8x22B-Instruct-v0.1`	—	66K	Frontier	Fair use (depends on server load)	Open
70	Llama 3.1 405B `meta-llama/Llama-3.1-405B`	7.3	131K	Frontier	Fair use (depends on server load)	Open
71	Llama 3.1 8B `nim/meta/llama-3.1-8b-instruct`	—	16K	Small	Fair use (depends on server load)	Open
72	Gemma 3 4B `google/gemma-3-4b-it`	—	66K	Small	Fair use (depends on server load)	Open
73	Llama 3.2 1B `meta-llama/Llama-3.2-1B`	—	131K	Small	Fair use (depends on server load)	Open
74	Mixtral 8x22b Instruct V01 `nim/mistralai/mixtral-8x22b-instruct-v01`	—	16K	Frontier	Fair use (depends on server load)	Open
75	Qwen3 4B `Qwen/Qwen3-4B-Instruct-2507`	—	262K	Small	Fair use (depends on server load)	Open
76	Molmo 7B D 0924 `allenai/Molmo-7B-D-0924`	5.6	4K	Small	Fair use (depends on server load)	Open
77	Gemma 3 270M `google/gemma-3-270m-it`	5.1	33K	Small	Fair use (depends on server load)	Open
78	Gemma 3 270M `google/gemma-3-270m-it-lora`	5.1	33K	Small	Fair use (depends on server load)	Open
79	DeepSeek R1 Distill Qwen 7B `deepseek-ai/DeepSeek-R1-Distill-Qwen-7B`	—	131K	Small	Fair use (depends on server load)	Open
80	Gemma 3 1B `google/gemma-3-1b-it`	—	33K	Small	Fair use (depends on server load)	Open
81	Cogito V1 Preview Llama 8B `deepcogito/cogito-v1-preview-llama-8B`	—	131K	Small	Fair use (depends on server load)	Open
82	Qwen2.5 3B `Qwen/Qwen2.5-3B-Instruct`	—	33K	Small	Fair use (depends on server load)	Open
83	Qwen2.5 1 5B `Qwen/Qwen2.5-1.5B-Instruct`	—	33K	Small	Fair use (depends on server load)	Open
84	Qwen2.5 7B `Qwen/Qwen2.5-7B`	—	131K	Small	Fair use (depends on server load)	Open
85	Qwen2.5 1 5B `Qwen/Qwen2.5-1.5B`	—	131K	Small	Fair use (depends on server load)	Open
86	Qwen2.5 7B `Qwen/Qwen2.5-7B-Instruct`	—	33K	Small	Fair use (depends on server load)	Open
87	Qwen3 8B `Qwen/Qwen3-8B`	—	41K	Small	Fair use (depends on server load)	Open
88	Qwen 3.0 6B `Qwen/Qwen3-0.6B`	—	41K	Small	Fair use (depends on server load)	Open
89	Qwen 3.1 7B `Qwen/Qwen3-1.7B`	—	41K	Small	Fair use (depends on server load)	Open
90	Meta Llama 3.1 8B `togethercomputer/meta-llama-3.1-8B-Instruct-AWQ-INT4`	—	131K	Small	Fair use (depends on server load)	Open
91	Sarvam M `sarvamai/sarvam-m`	—	33K	Small	Fair use (depends on server load)	Open
92	EssentialAI Rnj-1 Instruct `togethercomputer/EssentialAI-RNJ-1-Instruct`	—	33K	Small	Fair use (depends on server load)	Open
93	Qwen3 8B `Qwen/Qwen3-8B-Lora`	—	41K	Small	Fair use (depends on server load)	Open
94	Mistral 7B `mistralai/Mistral-7B-v0.1`	—	33K	Small	Fair use (depends on server load)	Open
95	Mixtral 8x7b `mistralai/Mixtral-8x7B-Instruct-v0.1-FP8-Lora`	—	33K	Mid-size	Fair use (depends on server load)	Open
96	Gemma 2B `google/gemma-2b-it`	—	8K	Small	Fair use (depends on server load)	Open
97	Llama 2 7B `meta-llama/Llama-2-7b-chat-hf`	—	4K	Small	Fair use (depends on server load)	Open
98	Gemma 2 9B `google/gemma-2-9b-it`	—	8K	Small	Fair use (depends on server load)	Open

Benchmarks: Artificial Analysis (October 24, 2026). A dash means the model is not on their leaderboard; we never estimate scores ourselves.

FAQ

Is Together AI free?

Only selected models are priced at zero; there is no general free trial. 98 models are currently free.

Is Together AI compatible with the OpenAI API?

Yes — use the base URL https://api.together.xyz/v1 with any OpenAI-compatible client or SDK.

Does Together AI require a card?

It depends on the account — check the provider's dashboard.

Together AI

FAQ

See also