TīrthaModels

Tīrtha · models you can use

Every model on Tīrtha. 68 and counting.

Models you can call through one API key, nearly all with open weights. 29 of them are tested on Tīrtha Bench so far; their cards show the numbers.

Start using Tīrtha in 3 steps

Read the docs
  1. Get an API keyFree to create. It works for every model in the catalogue.Get an API key
  2. Add creditsPay as you go. Credits work on every model.Add credits
  3. Point your client at TīrthaAny OpenAI-compatible client works. Use your key and this base URL.
    base_url: "https://api.tirtha.ai/v1" api_key: "YOUR_TIRTHA_KEY"

Qwen3.8-Max

by Qwen · Closed model, no public weightstext inimage in1M contextBench 96.1%

Qwen3-Next-80B-A3B-Instruct

by Qwen · Apache-2.0text in262K contextBench 95.6%

Qwen3-VL-235B-A22B-Instruct

by Qwen · Apache-2.0text inimage in262K contextBench 95.4%

Ling-3.1-flash

by inclusionAI · MITtext in1M contextBench 95.1%

Qwen3-30B-A3B

by Qwen · Apache-2.0text in40K contextBench 94.8%

gemma-4-31B-it (FP4)

by Google · Apache-2.0text inimage in262K contextBench 94.5%

DeepSeek-V4-Flash-0731

by DeepSeek · MITtext in1M contextBench 94.2%

GLM-4.7

by Z.ai · MITtext in202K contextBench 94.1%

Qwen3-Coder-480B-A35B-Instruct

by Qwen · Apache-2.0text in262K contextBench 94.1%

gemma-4-31B-it

by Google · Apache-2.0text inimage in262K contextBench 94.1%

DeepSeek-V4-Pro

by DeepSeek · MITtext in1M contextBench 94.1%

MiMo-V2.6-Pro

by Xiaomi · MITtext inimage inaudio invideo in1M contextBench 93.8%

GLM-5.3-Flash

by Z.ai · MITtext inimage in1M contextBench 93.6%

DeepSeek-V4.1-Flash

by DeepSeek · MITtext inimage in1M contextBench 93.6%

GLM-5.2

by Z.ai · MITtext in1M contextBench 93.6%

Llama-3.3-70B-Instruct

by Meta · Llama 3.3 Community Licensetext in131K contextBench 93.5%

Beam-1.0

by Reflection · Apache-2.0text inBench 92.8%

Qwen3.5-9B

by Qwen · Apache-2.0text inimage in262K contextBench 92.5%

NVIDIA-Nemotron-3.5-Super-VL

by NVIDIA · OpenMDW 1.1text inimage in262K contextBench 91.9%

Qwen3.8-Flash

by Qwen · Closed model, no public weightstext in1M contextBench 91.9%

gemma-3-27b-it

by Google · Gemma Terms of Usetext inimage in131K contextBench 90.8%

gemma-4-26B-A4B-it

by Google · Apache-2.0text inimage in262K contextBench 90.6%

Mistral-Small-3.2-24B-Instruct-2506

by Mistral AI · Apache-2.0text inimage in128K contextBench 90.2%

MiMo-V2.6-Flash

by Xiaomi · MITtext inimage inaudio invideo in1M contextBench 90.2%

Llama-3.1-8B-Instruct

by Meta · Llama 3.1 Community Licensetext in131K contextBench 80.5%

NVIDIA-Nemotron-3-Super-120B-A12B

by NVIDIA · NVIDIA Nemotron Open Model Licensetext in262K contextBench 72.5%

Muse-Glimmer-30B

by Meta Superintelligence Lab · Apache-2.0text inimage in131K contextBench 69.5%

gpt-oss-120b

by OpenAI · Apache-2.0text in131K contextBench 65.8%

DeepSeek-V3.2

by DeepSeek · MITtext in163K contextBench 53.8%

DeepSeek-V4-Pro-0813

by DeepSeek · MITtext in1M context

DeepSeek-V4.1-Flash-1017

by DeepSeek · MITtext inimage in1M context

DeepSeek-V4-Flash

by DeepSeek · MITtext in1M context

Qwen3-235B-A22B-Instruct-2507

by Qwen · Apache-2.0text in262K context

DeepSeek-R1-0528

by DeepSeek · MITtext in163K context

DeepSeek-V3

by DeepSeek · DeepSeek Model License (code: MIT)text in163K context

DeepSeek-V3.1

by DeepSeek · MITtext in163K context

DeepSeek-V4-Flash-Vision-Exp

by DeepSeek · MITtext inimage in1M context

gemma-3-4b-it

by Google · Gemma Terms of Usetext inimage in131K context

GLM-4.6

by Z.ai · MITtext in202K context

GLM-5.3

by Z.ai · GLM-5.3 Licensetext in1M context

gpt-oss-20b

by OpenAI · Apache-2.0text in131K context

granite-4.2-8b

by IBM · Apache-2.0text in524K context

Hy4-preview

by Tencent · Apache-2.0text in1M context

Inkling-Small

by Thinking Machines · Apache-2.0text inimage in524K context

Kimi-K2.6

by Moonshot AI · Modified MITtext inimage in262K context

Kimi-K3

by Moonshot AI · Kimi K3 Licensetext inimage in1M context

Ling-3.0-flash

by inclusionAI · MITtext in262K context

Ling-3.0-flash-VL

by inclusionAI · MITtext inimage in262K context

MiMo-V2.5-Pro

by Xiaomi · MITtext in

Ministral-3-14B-Instruct-2512

by Mistral AI · Apache-2.0text inimage in

Mistral-Nemo-Instruct-2407

by Mistral AI · Apache-2.0text in131K context

Mistral-Small-24B-Instruct-2501

by Mistral AI · Apache-2.0text in32K context

Nemotron-3-Nano-Omni

by NVIDIA · NVIDIA Open Model Agreementtext inimage in

Nemotron-Content-Safety-3.5

by NVIDIA · OpenMDW 1.1text in131K context

NVIDIA-Nemotron-3-Ultra-550B-A55B

by NVIDIA · OpenMDW 1.1text in262K context

NVIDIA-Nemotron-3.5-Lightning

by NVIDIA · OpenMDW 1.1text in262K context

NVIDIA-Nemotron-Nano-12B-v2-VL

by NVIDIA · NVIDIA Open Model Licensetext inimage in

phi-4

by Microsoft · MITtext in16K context

Qwen2.5-72B-Instruct

by Qwen · Qwen Licensetext in131K context

Qwen3-14B

by Qwen · Apache-2.0text in40K context

Qwen3-32B

by Qwen · Apache-2.0text in40K context

Qwen3-VL-30B-A3B-Instruct

by Qwen · Apache-2.0text inimage in262K context

Qwen3.5-27B

by Qwen · Apache-2.0text inimage in262K context

Qwen3.5-35B-A3B

by Qwen · Apache-2.0text inimage in262K context

Qwen3.5-397B-A17B

by Qwen · Apache-2.0text inimage in262K context

Qwen3.6-27B

by Qwen · Apache-2.0text inimage in262K context

Qwen3.6-35B-A3B

by Qwen · Apache-2.0text inimage in262K context

Qwen3.8-27B

by Qwen · Apache-2.0text inimage in262K context