Together AI Publishes DeepSeek·GPT-5.6 Luna Cascade Test Results
DeepSeek-first cascade solves more tasks than Luna alone on DeepSWE benchmark
00
Together AI
An AI cloud that runs open-weight models on behalf of others, serving as the server room for companies wanting to use open-source models. Its flagship service is offering DeepSeek's latest model, V4 Pro, as an API, and it published its pricing sheet alongside the full rollout in August 2026. That same month, it expanded its infrastructure by supplying 10,000 NVIDIA B300 GPUs to India's largest AI data center. It was also revealed in August 2026 that, together with IBM and NVIDIA, it had built a B300 cluster on IBM Cloud. Around the same time, news emerged that Yutori's browser agent runs twice as fast as before on Together AI's environment.
Current rank (1M)
28
- Change
- Prev.
- 15
Rank over the last 30 days
2026.08.17 – 2026.09.15 · High 14 · Low 28 · Now 28
Together AI · Related stories
That's the last story.
Together AI · Recent announcements
- 640glm-5.3 now beats gpt-5.6 sol and claude fable 5 on agentic benchmarks
- 614an underrated part of glm-5.3 is its low hallucination rate
- 481We just signed one of the largest AI infrastructure deals for open source, period.
- 190GLM-5.3 Flash has arrived.
- 143Deciding should you use GLM-5.3 or GLM-5.3 Flash?
- 131When should you use GLM-5.3 vs. GLM-5.3 Flash?
- 108GLM-5.3 has arrived on Together AI on Day 0
- 90Top open models by use case 👇
- 86for september, we’re cutting the price of Dedicated Inference on H100s from $5.49/hr to $3.99/hr
- 76Qwen3.8 27B is now available for both fine-tuning and Dedicated Model Inference
- 71Serving Kimi K3 at scale: An inference deep dive with Together AI
- 32Excited to welcome @greptile, the AI code reviewer trusted by 11,000+ teams, as a Together AI customer.
- 24Proud to continue to partner with @yutori_ai to power their Navigator product suite!
- 23Compute has always been core to the @VerneRobotics story.
- 16We're honored to be named a 2026 #IA40 winner.
- 6🔜 https://t.co/7UYAtgUP6q
- –GLM-5.3 vs. GLM-5.3 Flash on DeepSWE: Cost, Coding, and Routing