{"schema_version":"onlylabs.public_analysis.v1","url":"https://onlylabs.fyi/analysis/stepfun","json_url":"https://onlylabs.fyi/analysis/stepfun/analysis.json","evidence_json_url":"https://onlylabs.fyi/analysis/stepfun/evidence.json","generated_at":"2026-06-28T02:30:42.529Z","analysis":{"org_slug":"stepfun","url":"https://onlylabs.fyi/analysis/stepfun","json_url":"https://onlylabs.fyi/analysis/stepfun/analysis.json","evidence_json_url":"https://onlylabs.fyi/analysis/stepfun/evidence.json","dossier_url":"https://onlylabs.fyi/labs/stepfun","org":{"slug":"stepfun","name":"StepFun","category":"neolab","category_label":"Neolab","homepage_url":"https://www.stepfun.com"},"title":"StepFun analysis","summary":"StepFun is a Shanghai-based frontier AI lab executing an unusually broad multimodal strategy—spanning video generation, real-time voice, image editing, 3D asset generation, formal mathematics, GUI agents, and deep research—while releasing the vast majority of its model weights, code, and benchmarks under permissive open-source licenses. The lab's May 2026 release of Step 3.7 Flash, a 198B MoE vision-language model…","markdown":"## Thesis\n\nStepFun is a Shanghai-based frontier AI lab executing an unusually broad multimodal strategy—spanning video generation, real-time voice, image editing, 3D asset generation, formal mathematics, GUI agents, and deep research—while releasing the vast majority of its model weights, code, and benchmarks under permissive open-source licenses. The lab's May 2026 release of Step 3.7 Flash, a 198B MoE vision-language model with native image/video input and 256K context [E3](https://huggingface.co/stepfun-ai/Step-3.7-Flash)[W1](https://static.stepfun.com/blog/step-3.7-flash/)[W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/), paired with the June 2026 launch of a voice-driven CLI coding assistant [P1](https://github.com/stepfun-ai/Step-Realtime-CLI)[P2](https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0), signals a push toward agentic and real-time multimodal product surfaces. StepFun's pace of open-weight releases—spanning 10B to 321B parameters across text, vision, audio, and image generation—positions it as one of the most prolific open-source contributors among frontier labs, with commercial monetization routed through its API platform and NVIDIA NIM distribution [W1](https://static.stepfun.com/blog/step-3.7-flash/)[W2](http://github.com/stepfun-ai/Step-3.7-Flash)[P6](https://github.com/stepfun-ai/Step-Realtime-Console).\n\n## Signal desks\n\n### Hiring\n\nNo cited evidence in this pack.\n\n### Forks\n\n- **vllm** — Forked from `vllm-project/vllm` (Feb 2025). The fork contains a `step-audio` branch, indicating StepFun maintains a customized vLLM backend for its audio models. 45 stars. [E53](https://github.com/stepfun-ai/vllm)[P12](https://github.com/stepfun-ai/Step-Audio2)\n- **llama.cpp** — Forked from `ggml-org/llama.cpp` (Mar 2026). Supports GGUF quantization serving for Step 3.7 Flash, suggesting investment in local/edge inference compatibility. 6 stars. [E60](https://github.com/stepfun-ai/llama.cpp)[W2](http://github.com/stepfun-ai/Step-3.7-Flash)\n\n### Releases\n\n- **Step 3.7 Flash** — May 2026. 198B MoE vision-language model (201B params), Apache 2.0. BF16, FP8, NVFP4, GGUF checkpoints. 142K HF downloads, 392 likes. Native image/video input, 256K context, three reasoning levels. Distributed as NVIDIA NIM microservice. [E3](https://huggingface.co/stepfun-ai/Step-3.7-Flash)[W1](https://static.stepfun.com/blog/step-3.7-flash/)[W2](http://github.com/stepfun-ai/Step-3.7-Flash)[W3](https://www.marktechpost.com/2026/05/29/stepfun-releases-step-3-7-flash-a-198b-moe-vision-language-model-for-coding-agents-and-search-workflows/)[W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/)\n- **Step 3.5 Flash** — Feb 2026. Text-generation model (~199B params). 224K HF downloads, 820 likes. Plus base and midtrain checkpoints released Mar 2026. [E1](https://huggingface.co/stepfun-ai/Step-3.5-Flash)[E10](https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base)[E16](https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base-Midtrain)\n- **Step3** — Jul 2025. 321B MoE multimodal reasoning model (38B active). 156K HF downloads, 166 likes. [E5](https://huggingface.co/stepfun-ai/step3)[P14](https://github.com/stepfun-ai/Step3)\n- **Step3-VL-10B** — Jan 2026. Compact 10B multimodal model, 469K HF downloads, 410 likes. FP8 variant also released. [E2](https://huggingface.co/stepfun-ai/Step3-VL-10B)[E12](https://huggingface.co/stepfun-ai/Step3-VL-10B-Base)[E35](https://huggingface.co/stepfun-ai/Step3-VL-10B-FP8)[P28](https://github.com/stepfun-ai/Step3-VL-10B)\n- **Step-Audio-R1 / R1.1 / R1.5** — Nov 2025–Apr 2026. Speech reasoning models (33B params). Apache 2.0. [E4](https://huggingface.co/stepfun-ai/Step-Audio-R1.1)[E7](https://huggingface.co/stepfun-ai/Step-Audio-R1)[P19](https://github.com/stepfun-ai/Step-Audio-R1)\n- **Step-Audio-EditX** — Oct 2025. 3B-parameter RL-based audio editing model. 929 GitHub stars. Multi-language (JP, KR added Nov 2025). [E41](https://github.com/stepfun-ai/Step-Audio-EditX)[P18](https://github.com/stepfun-ai/Step-Audio-EditX)\n- **Step-Audio2 / mini / mini-Think** — Jul–Sep 2025. End-to-end multimodal LLM for audio understanding and speech conversation. 1,460 GitHub stars. [E22](https://github.com/stepfun-ai/Step-Audio2)[P12](https://github.com/stepfun-ai/Step-Audio2)\n- **Step-Video-T2V / Turbo / TI2V** — Feb–Mar 2025. Text-to-video and image-to-video models. T2V: 3,185 stars. TI2V: 375 stars. Both with model weights and tech reports. [E18](https://github.com/stepfun-ai/Step-Video-T2V)[E49](https://github.com/stepfun-ai/Step-Video-TI2V)[P3](https://github.com/stepfun-ai/Step-Video-T2V)[P4](https://github.com/stepfun-ai/Step-Video-TI2V)\n- **Step1X-Edit / v1p2** — Apr 2025–Apr 2026. SOTA open-source image editing. v1: 2,224 stars. v1p2: HF release with diffusers support. Step Image Edit 2 released Apr 2026 (API-only, sub-2s generation). [E23](https://github.com/stepfun-ai/Step1X-Edit)[E11](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2)[E25](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2-preview)[E40](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p1-diffusers)[P8](https://github.com/stepfun-ai/Step1X-Edit)\n- **NextStep-1 / 1.1 / Large** — Aug 2025–Feb 2026. Autoregressive image generation with continuous tokens (14B params). ICLR 2026 Oral. 689 stars. [E9](https://huggingface.co/stepfun-ai/NextStep-1-Large)[E13](https://huggingface.co/stepfun-ai/NextStep-1-Large-Edit)[E21](https://huggingface.co/stepfun-ai/NextStep-1.1)[E24](https://huggingface.co/stepfun-ai/NextStep-1-Large-Pretrain)[E27](https://huggingface.co/stepfun-ai/NextStep-1-f8ch16-Tokenizer)[E28](https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain-256px)[E38](https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain)[P15](https://github.com/stepfun-ai/NextStep-1)\n- **Step1X-3D** — May 2025. High-fidelity textured 3D asset generation. 869 stars. [E8](https://huggingface.co/stepfun-ai/Step1X-3D)[E42](https://github.com/stepfun-ai/Step1X-3D)[P10](https://github.com/stepfun-ai/Step1X-3D)\n- **GELab-Zero-4B-preview** — Nov 2025. Open-source GUI agent with model and infrastructure. 2,191 GitHub stars, MCP server support. [E6](https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview)[E31](https://github.com/stepfun-ai/gelab-zero)[P22](https://github.com/stepfun-ai/gelab-zero)\n- **StepDeepResearch** — Nov 2025. Cost-effective end-to-end deep research agent. 561 stars. Latest: Step 3.5 Flash achieves 65.27 on ResearchRubrics. [E46](https://github.com/stepfun-ai/StepDeepResearch)[P23](https://github.com/stepfun-ai/StepDeepResearch)\n- **PaCoRe-8B / RLVR-8B** — Dec 2025. Parallel coordinated reasoning framework. 94.5% on HMMT 2025, surpassing GPT-5's 93.2%. MIT license. 334 stars. [E15](https://huggingface.co/stepfun-ai/PaCoRe-8B)[E30](https://github.com/stepfun-ai/PaCoRe)[E36](https://huggingface.co/stepfun-ai/RLVR-8B-0926)[P24](https://github.com/stepfun-ai/PaCoRe)\n- **StepFun-Prover-Preview (7B/32B)** — Aug 2025. Formal theorem proving in Lean 4. 70% pass@1 on miniF2F-test (32B). [E33](https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-32B)[E45](https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-7B)[P16](https://github.com/stepfun-ai/StepFun-Prover-Preview)\n- **StepFun-Formalizer (7B/32B)** — Aug 2025. Autoformalization to Lean 4. [E32](https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B)[E39](https://huggingface.co/stepfun-ai/StepFun-Formalizer-7B)[P17](https://github.com/stepfun-ai/StepFun-Formalizer)\n- **StepMesh** — Jul 2025. High-performance C++ communication library for Attention-FFN disaggregation. 367 stars. [E50](https://github.com/stepfun-ai/StepMesh)[P13](https://github.com/stepfun-ai/StepMesh)\n- **SteptronOss** — Dec 2025. Lightweight training framework for SFT, RLVR, and evaluation. 575 stars. [E37](https://github.com/stepfun-ai/SteptronOss)[P27](https://github.com/stepfun-ai/SteptronOss)\n- **Step-Realtime-CLI v0.1.0** — Jun 2026. Terminal-based AI coding assistant with real-time voice. TypeScript, MIT. [E17](https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0)[P1](https://github.com/stepfun-ai/Step-Realtime-CLI)[P2](https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0)\n- **Step-Realtime-Console** — Mar 2025. Svelte-based real-time voice API demo with WaveSurfer visualization. 74 stars. [E51](https://github.com/stepfun-ai/Step-Realtime-Console)[P6](https://github.com/stepfun-ai/Step-Realtime-Console)\n- **ComfyUI-StepVideo** — Mar 2025. Custom ComfyUI nodes for StepVideo TI2V/T2V. 43 stars. [E54](https://github.com/stepfun-ai/ComfyUI-StepVideo)[P7](https://github.com/stepfun-ai/ComfyUI-StepVideo)\n- **Step-Audio-Edit-Benchmark** — Nov 2025. Evaluation framework for controllable speech synthesis. 20 stars. [E58](https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark)[P21](https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark)\n- **InfiniteHBD-Trace** — May 2025. SIGCOMM 2025 fault-trace dataset from 400 GPU servers for LLM pretraining. [E59](https://github.com/stepfun-ai/InfiniteHBD-Trace)[P11](https://github.com/stepfun-ai/InfiniteHBD-Trace)\n- **GEBench** — Feb 2026. (Python repo, 54 stars). [E52](https://github.com/stepfun-ai/GEBench)\n- **StepAudio-Skills** — Mar 2026. Audio skills for Claw. [E57](https://github.com/stepfun-ai/StepAudio-Skills)\n- **stepfunApi-audio-sdk** — Nov 2025. Android audio SDK (Kotlin) for TTS/ASR. [P20](https://github.com/stepfun-ai/stepfunApi-audio-sdk)\n- **Qwen2.5-32B-DialogueReason** — May 2025. Dialogue reasoning model based on Qwen2.5. [E29](https://huggingface.co/stepfun-ai/Qwen2.5-32B-DialogueReason)\n\n### Talking\n\n- **Step 3.7 Flash launch narrative**: Blog post and GitHub README emphasize enterprise agentic workflows—\"Fast, Sharp & Reliable Agentic Intelligence\"—with vision input, tool-use reliability, and multi-level reasoning. Ecosystem support spans vLLM, SGLang, llama.cpp, NVIDIA Nemo (AutoModel, Megatron Core, Megatron Bridge), and NVIDIA NIM. [W1](https://static.stepfun.com/blog/step-3.7-flash/)[W2](http://github.com/stepfun-ai/Step-3.7-Flash)[W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/)\n- **Third-party coverage (MarkTechPost)**: Frames Step 3.7 Flash as targeting \"coding agents and search workflows,\" noting the multimodal MoE architecture and Apache 2.0 licensing with BF16/FP8/NVFP4/GGUF weights. [W3](https://www.marktechpost.com/2026/05/29/stepfun-releases-step-3-7-flash-a-198b-moe-vision-language-model-for-coding-agents-and-search-workflows/)\n- **Third-party coverage (NVIDIA Technical Blog)**: Highlights enterprise readiness, NVFP4 quantization for reduced memory bandwidth, and 256K context window for financial analysis and concurrent coding agents. [W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/)\n- **Voice AI narrative (Decrypt)**: Covers StepAudio 2.5 Realtime as an end-to-end voice model (audio-in, audio-out, no text conversion) with customizable personas and a \"soul-level companion\" named Xiao Yue. Positions StepFun as competing on voice benchmarks. [W5](https://decrypt.co/369013/stepfun-stepaudio-voice-ai-tops-benchmarks)\n- **Developer tool framing (Step-Realtime-CLI)**: Positions the CLI as a voice-driven coding assistant—`step voice` for spoken code editing, `step exec --mode plan` for read-only planning. Dual-region deployment (Mainland China and Overseas) with separate API endpoints. [P1](https://github.com/stepfun-ai/Step-Realtime-CLI)[W6](https://p.rst.im/q/github.com/stepfun-ai/Step-Realtime-CLI)\n- **Academic recognition**: NextStep-1 received ICLR 2026 Oral [P15](https://github.com/stepfun-ai/NextStep-1); InfiniteHBD-Trace accepted at ACM SIGCOMM 2025 [P11](https://github.com/stepfun-ai/InfiniteHBD-Trace); multiple arXiv tech reports accompany model releases [P3](https://github.com/stepfun-ai/Step-Video-T2V)[P4](https://github.com/stepfun-ai/Step-Video-TI2V)[P5](https://github.com/stepfun-ai/Step-Audio)[P8](https://github.com/stepfun-ai/Step1X-Edit)[P10](https://github.com/stepfun-ai/Step1X-3D)[P12](https://github.com/stepfun-ai/Step-Audio2)[P14](https://github.com/stepfun-ai/Step3)[P15](https://github.com/stepfun-ai/NextStep-1)[P16](https://github.com/stepfun-ai/StepFun-Prover-Preview)[P17](https://github.com/stepfun-ai/StepFun-Formalizer)[P18](https://github.com/stepfun-ai/Step-Audio-EditX)[P19](https://github.com/stepfun-ai/Step-Audio-R1)[P23](https://github.com/stepfun-ai/StepDeepResearch)[P24](https://github.com/stepfun-ai/PaCoRe)[P28](https://github.com/stepfun-ai/Step3-VL-10B).\n\n## Shipping\n\nStepFun's shipping cadence is unusually high and spans modalities. Key shipped artifacts with inspectable evidence:\n\n- **Models on Hugging Face**: 30+ distinct model repositories released since Feb 2025, all under Apache 2.0 or MIT licenses, covering text (Step3 family: 321B→3.5 Flash→3.7 Flash), vision-language (Step3-VL-10B), audio (Step-Audio, Audio2, Audio-R1, Audio-EditX), video (T2V, TI2V), image generation (NextStep-1, Step1X-Edit, Step1X-3D), reasoning (PaCoRe), and agents (GELab-Zero). [E21](https://huggingface.co/stepfun-ai/NextStep-1.1)[E24](https://huggingface.co/stepfun-ai/NextStep-1-Large-Pretrain)[E25](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2-preview)[E32](https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B)[E33](https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-32B)[E35](https://huggingface.co/stepfun-ai/Step3-VL-10B-FP8)[E36](https://huggingface.co/stepfun-ai/RLVR-8B-0926)[E45](https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-7B)\n- **Infrastructure software**: StepMesh (C++ communication library for attention-FFN disaggregation) [P13](https://github.com/stepfun-ai/StepMesh); SteptronOss (training framework for SFT/RLVR/eval) [P27](https://github.com/stepfun-ai/SteptronOss); Step-Realtime-CLI (voice coding tool) [P1](https://github.com/stepfun-ai/Step-Realtime-CLI)[P2](https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0); Step-Realtime-Console (voice API Svelte demo) [P6](https://github.com/stepfun-ai/Step-Realtime-Console); ComfyUI-StepVideo (custom nodes) [P7](https://github.com/stepfun-ai/ComfyUI-StepVideo); Android audio SDK [P20](https://github.com/stepfun-ai/stepfunApi-audio-sdk).\n- **Datasets and benchmarks**: AndroidDaily (GUI agent dataset) [P22](https://github.com/stepfun-ai/gelab-zero); Step-Audio-Edit-Benchmark [P21](https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark); Step-Video-TI2V-Eval [P4](https://github.com/stepfun-ai/Step-Video-TI2V); StepEval-Audio-360 [P5](https://github.com/stepfun-ai/Step-Audio); GEdit-Bench [P8](https://github.com/stepfun-ai/Step1X-Edit); InfiniteHBD-Trace (SIGCOMM 2025 fault traces) [P11](https://github.com/stepfun-ai/InfiniteHBD-Trace); PaCoRe-Train-8k [P24](https://github.com/stepfun-ai/PaCoRe); three in-house benchmarks from Step-Audio-R1.5 [P19](https://github.com/stepfun-ai/Step-Audio-R1).\n- **Platform**: API available at both platform.stepfun.com (Mainland China) and platform.stepfun.ai (Overseas). Models served via API with NVIDIA NIM distribution for enterprise. [P1](https://github.com/stepfun-ai/Step-Realtime-CLI)[P6](https://github.com/stepfun-ai/Step-Realtime-Console)[P8](https://github.com/stepfun-ai/Step1X-Edit)[P23](https://github.com/stepfun-ai/StepDeepResearch)[W1](https://static.stepfun.com/blog/step-3.7-flash/)[W2](http://github.com/stepfun-ai/Step-3.7-Flash)\n- **Notable gaps**: No cited evidence of mobile app releases, chat products, or consumer-facing products beyond the API platform and developer tools.\n\n## Research themes\n\nStepFun's research agenda clusters around five interconnected themes:\n\n1. **Multimodal foundation models at scale**: The Step3 family demonstrates a progression from a 321B MoE (38B active) [P14](https://github.com/stepfun-ai/Step3) to the 198B Step 3.7 Flash with native vision [E3](https://huggingface.co/stepfun-ai/Step-3.7-Flash)[W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/) and the compact 10B Step3-VL-10B that \"matches open-source models 10-20x its size\" [P28](https://github.com/stepfun-ai/Step3-VL-10B). All use MoE architectures optimized for inference cost.\n\n2. **Real-time speech interaction**: A sustained investment from Step-Audio (Feb 2025, now deprecated) [P5](https://github.com/stepfun-ai/Step-Audio) through Step-Audio2 (end-to-end multimodal LLM for speech) [P12](https://github.com/stepfun-ai/Step-Audio2), Step-Audio-R1/R1.5 (speech reasoning with RL) [P19](https://github.com/stepfun-ai/Step-Audio-R1), to Step-Audio-EditX (3B RL-based audio editing with emotion/paralinguistic control) [P18](https://github.com/stepfun-ai/Step-Audio-EditX). The Realtime Console [P6](https://github.com/stepfun-ai/Step-Realtime-Console), Realtime CLI [P1](https://github.com/stepfun-ai/Step-Realtime-CLI), and Android SDK [P20](https://github.com/stepfun-ai/stepfunApi-audio-sdk) form the application layer.\n\n3. **Visual content generation**: Separate tracks for video (T2V diffusion, TI2V with motion control) [P3](https://github.com/stepfun-ai/Step-Video-T2V)[P4](https://github.com/stepfun-ai/Step-Video-TI2V), image editing (Step1X-Edit targeting GPT-4o/Gemini parity) [P8](https://github.com/stepfun-ai/Step1X-Edit), autoregressive image generation (NextStep-1 with continuous tokens, ICLR 2026 Oral) [P15](https://github.com/stepfun-ai/NextStep-1), and 3D asset generation (Step1X-3D) [P10](https://github.com/stepfun-ai/Step1X-3D).\n\n4. **Reasoning and formal methods**: PaCoRe introduces parallel coordinated reasoning that scales test-time compute to ~2M tokens, achieving 94.5% on HMMT 2025 (surpassing GPT-5) [P24](https://github.com/stepfun-ai/PaCoRe). StepFun-Prover-Preview achieves 70% pass@1 on miniF2F-test via tool-integrated RL with Lean 4 [P16](https://github.com/stepfun-ai/StepFun-Prover-Preview). StepFun-Formalizer tackles autoformalization [P17](https://github.com/stepfun-ai/StepFun-Formalizer). StepDeepResearch applies agentic reasoning to research tasks [P23](https://github.com/stepfun-ai/StepDeepResearch).\n\n5. **Agent infrastructure**: GELab-Zero provides a fully open-source GUI agent (model + infrastructure, MCP server support, \"no cloud dependencies\") [P22](https://github.com/stepfun-ai/gelab-zero). Step-Realtime-CLI extends agent capabilities to voice-driven coding [P1](https://github.com/stepfun-ai/Step-Realtime-CLI). SteptronOss supports the RLVR training pipeline that underlies reasoning models [P27](https://github.com/stepfun-ai/SteptronOss). StepMesh addresses the infrastructure challenge of attention-FFN disaggregation for serving large models [P13](https://github.com/stepfun-ai/StepMesh).\n\n## Hiring & scaling\n\nNo cited evidence in this pack. No open job listings, team descriptions, location expansions, or hiring signals appear in any of the provided sources.\n\n## Category implications\n\n**Strategy**: StepFun is executing a broad-coverage open-source strategy that releases model weights, code, benchmarks, and training frameworks under Apache 2.0/MIT while monetizing through API access (platform.stepfun.com / platform.stepfun.ai) and NVIDIA NIM enterprise distribution [W1](https://static.stepfun.com/blog/step-3.7-flash/)[W2](http://github.com/stepfun-ai/Step-3.7-Flash)[W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/). The dual-region operational split (Mainland China vs. Overseas with separate accounts and API endpoints) [P1](https://github.com/stepfun-ai/Step-Realtime-CLI) suggests deliberate compliance architecture for cross-border deployment.\n\n**Infrastructure**: The StepMesh library for attention-FFN disaggregation [P13](https://github.com/stepfun-ai/StepMesh), the SteptronOss training framework [P27](https://github.com/stepfun-ai/SteptronOss), forks of vLLM and llama.cpp [E53](https://github.com/stepfun-ai/vllm)[E60](https://github.com/stepfun-ai/llama.cpp), and NVIDIA Nemo ecosystem integration (AutoModel, Megatron Core, Megatron Bridge) [W1](https://static.stepfun.com/blog/step-3.7-flash/) collectively indicate heavy investment in both training and inference infrastructure. The InfiniteHBD-Trace dataset (SIGCOMM 2025) reveals operation of GPU clusters with hundreds of nodes for LLM pretraining [P11](https://github.com/stepfun-ai/InfiniteHBD-Trace).\n\n**Product**: The Step-Realtime-CLI release signals a move toward developer-facing voice-coding tools as a product surface [P1](https://github.com/stepfun-ai/Step-Realtime-CLI)[P2](https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0). The Realtime Console [P6](https://github.com/stepfun-ai/Step-Realtime-Console) and Android SDK [P20](https://github.com/stepfun-ai/stepfunApi-audio-sdk) suggest platform ambitions in real-time voice. The GELab-Zero GUI agent with MCP server support targets multi-device agent orchestration [P22](https://github.com/stepfun-ai/gelab-zero). StepDeepResearch targets the research agent market [P23](https://github.com/stepfun-ai/StepDeepResearch).\n\n**Research**: The breadth—spanning video, audio, images, 3D, code, formal math, and agents—is unusual for a single lab and suggests a \"multimodal everything\" research thesis. The ICLR 2026 Oral (NextStep-1) [P15](https://github.com/stepfun-ai/NextStep-1) and SIGCOMM 2025 (InfiniteHBD-Trace) [P11](https://github.com/stepfun-ai/InfiniteHBD-Trace) acceptances indicate academic credibility. The consistent release of tech reports alongside model weights on arXiv [P3](https://github.com/stepfun-ai/Step-Video-T2V)[P4](https://github.com/stepfun-ai/Step-Video-TI2V)[P5](https://github.com/stepfun-ai/Step-Audio)[P8](https://github.com/stepfun-ai/Step1X-Edit)[P10](https://github.com/stepfun-ai/Step1X-3D)[P12](https://github.com/stepfun-ai/Step-Audio2)[P14](https://github.com/stepfun-ai/Step3)[P15](https://github.com/stepfun-ai/NextStep-1)[P16](https://github.com/stepfun-ai/StepFun-Prover-Preview)[P17](https://github.com/stepfun-ai/StepFun-Formalizer)[P18](https://github.com/stepfun-ai/Step-Audio-EditX)[P19](https://github.com/stepfun-ai/Step-Audio-R1)[P23](https://github.com/stepfun-ai/StepDeepResearch)[P24](https://github.com/stepfun-ai/PaCoRe)[P28](https://github.com/stepfun-ai/Step3-VL-10B) suggests a deliberate research communication strategy.\n\n**GTM**: The lab reaches developers through GitHub (cumulative stars exceeding 15,000 across repos), Hugging Face (30+ model repositories), academic papers, and NVIDIA's enterprise channel. The third-party coverage from Decrypt [W5](https://decrypt.co/369013/stepfun-stepaudio-voice-ai-tops-benchmarks), MarkTechPost [W3](https://www.marktechpost.com/2026/05/29/stepfun-releases-step-3-7-flash-a-198b-moe-vision-language-model-for-coding-agents-and-search-workflows/), and NVIDIA's blog [W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/) suggests an active media relations function. The WeChat and Discord community channels [P8](https://github.com/stepfun-ai/Step1X-Edit)[P12](https://github.com/stepfun-ai/Step-Audio2)[P23](https://github.com/stepfun-ai/StepDeepResearch) indicate developer community building.\n\n## Traction highlights\n\n- **GitHub stars**: Step-Video-T2V: 3,185 [P3](https://github.com/stepfun-ai/Step-Video-T2V); Step1X-Edit: 2,224 [P8](https://github.com/stepfun-ai/Step1X-Edit); GELab-Zero: 2,191 [P22](https://github.com/stepfun-ai/gelab-zero); Step 3.5 Flash repo: 2,089 [E34](https://github.com/stepfun-ai/Step-3.5-Flash); Step-Audio2: 1,460 [P12](https://github.com/stepfun-ai/Step-Audio2); Step-Audio-EditX: 929 [P18](https://github.com/stepfun-ai/Step-Audio-EditX); Step1X-3D: 869 [P10](https://github.com/stepfun-ai/Step1X-3D); NextStep-1: 689 [P15](https://github.com/stepfun-ai/NextStep-1); Step-Audio-R1: 673 [P19](https://github.com/stepfun-ai/Step-Audio-R1); SteptronOss: 575 [P27](https://github.com/stepfun-ai/SteptronOss); StepDeepResearch: 561 [P23](https://github.com/stepfun-ai/StepDeepResearch); Step3: 453 [P14](https://github.com/stepfun-ai/Step3); Step3-VL-10B: 407 [P28](https://github.com/stepfun-ai/Step3-VL-10B); Step-Video-TI2V: 375 [P4](https://github.com/stepfun-ai/Step-Video-TI2V); StepMesh: 367 [P13](https://github.com/stepfun-ai/StepMesh); PaCoRe: 334 [P24](https://github.com/stepfun-ai/PaCoRe); Step 3.7 Flash repo: 253 [E20](https://github.com/stepfun-ai/Step-3.7-Flash)\n- **HF downloads**: Step3-VL-10B: 469K [E2](https://huggingface.co/stepfun-ai/Step3-VL-10B); Step 3.5 Flash: 224K [E1](https://huggingface.co/stepfun-ai/Step-3.5-Flash); Step3: 156K [E5](https://huggingface.co/stepfun-ai/step3); Step 3.7 Flash: 142K [E3](https://huggingface.co/stepfun-ai/Step-3.7-Flash); NextStep-1.1: 6,222 [E21](https://huggingface.co/stepfun-ai/NextStep-1.1); GELab-Zero-4B: 639 [E6](https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview); Step1X-Edit-v1p2: 507 [E11](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2); Step-Audio-R1: 435 [E7](https://huggingface.co/stepfun-ai/Step-Audio-R1); Step-Audio-R1.1: 260 [E4](https://huggingface.co/stepfun-ai/Step-Audio-R1.1); Step3-VL-10B-FP8: 198 [E35](https://huggingface.co/stepfun-ai/Step3-VL-10B-FP8); StepFun-Formalizer-7B: 168 [E39](https://huggingface.co/stepfun-ai/StepFun-Formalizer-7B); Step1X-Edit-v1p1-diffusers: 161 [E40](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p1-diffusers); StepFun-Formalizer-32B: 128 [E32](https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B)\n- **Academic recognition**: ICLR 2026 Oral (NextStep-1) [P15](https://github.com/stepfun-ai/NextStep-1); SIGCOMM 2025 (InfiniteHBD-Trace) [P11](https://github.com/stepfun-ai/InfiniteHBD-Trace)\n- **Press coverage**: Decrypt [W5](https://decrypt.co/369013/stepfun-stepaudio-voice-ai-tops-benchmarks), MarkTechPost [W3](https://www.marktechpost.com/2026/05/29/stepfun-releases-step-3-7-flash-a-198b-moe-vision-language-model-for-coding-agents-and-search-workflows/), NVIDIA Technical Blog [W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/)\n\n## Sources\n\n[P1](https://github.com/stepfun-ai/Step-Realtime-CLI) stepfun-ai/Step-Realtime-CLI repository metadata\n[P2](https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0) stepfun-ai/Step-Realtime-CLI v0.1.0 release\n[P3](https://github.com/stepfun-ai/Step-Video-T2V) stepfun-ai/Step-Video-T2V repository metadata\n[P4](https://github.com/stepfun-ai/Step-Video-TI2V) stepfun-ai/Step-Video-TI2V repository metadata\n[P5](https://github.com/stepfun-ai/Step-Audio) stepfun-ai/Step-Audio repository metadata\n[P6](https://github.com/stepfun-ai/Step-Realtime-Console) stepfun-ai/Step-Realtime-Console repository metadata\n[P7](https://github.com/stepfun-ai/ComfyUI-StepVideo) stepfun-ai/ComfyUI-StepVideo repository metadata\n[P8](https://github.com/stepfun-ai/Step1X-Edit) stepfun-ai/Step1X-Edit repository metadata\n[P9](https://github.com/stepfun-ai/Step-Realtime-Python-Demo) stepfun-ai/Step-Realtime-Python-Demo repository metadata\n[P10](https://github.com/stepfun-ai/Step1X-3D) stepfun-ai/Step1X-3D repository metadata\n[P11](https://github.com/stepfun-ai/InfiniteHBD-Trace) stepfun-ai/InfiniteHBD-Trace repository metadata\n[P12](https://github.com/stepfun-ai/Step-Audio2) stepfun-ai/Step-Audio2 repository metadata\n[P13](https://github.com/stepfun-ai/StepMesh) stepfun-ai/StepMesh repository metadata\n[P14](https://github.com/stepfun-ai/Step3) stepfun-ai/Step3 repository metadata\n[P15](https://github.com/stepfun-ai/NextStep-1) stepfun-ai/NextStep-1 repository metadata\n[P16](https://github.com/stepfun-ai/StepFun-Prover-Preview) stepfun-ai/StepFun-Prover-Preview repository metadata\n[P17](https://github.com/stepfun-ai/StepFun-Formalizer) stepfun-ai/StepFun-Formalizer repository metadata\n[P18](https://github.com/stepfun-ai/Step-Audio-EditX) stepfun-ai/Step-Audio-EditX repository metadata\n[P19](https://github.com/stepfun-ai/Step-Audio-R1) stepfun-ai/Step-Audio-R1 repository metadata\n[P20](https://github.com/stepfun-ai/stepfunApi-audio-sdk) stepfun-ai/stepfunApi-audio-sdk repository metadata\n[P21](https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark) stepfun-ai/Step-Audio-Edit-Benchmark repository metadata\n[P22](https://github.com/stepfun-ai/gelab-zero) stepfun-ai/gelab-zero repository metadata\n[P23](https://github.com/stepfun-ai/StepDeepResearch) stepfun-ai/StepDeepResearch repository metadata\n[P24](https://github.com/stepfun-ai/PaCoRe) stepfun-ai/PaCoRe repository metadata\n[P25](https://github.com/stepfun-ai/Step1X-3D/releases/tag/v.1.0.0) stepfun-ai/Step1X-3D v.1.0.0 release\n[P26](https://github.com/stepfun-ai/Step1X-3D/releases/tag/v.1.0.1) stepfun-ai/Step1X-3D v.1.0.1 release\n[P27](https://github.com/stepfun-ai/SteptronOss) stepfun-ai/SteptronOss repository metadata\n[P28](https://github.com/stepfun-ai/Step3-VL-10B) stepfun-ai/Step3-VL-10B repository metadata\n[E1](https://huggingface.co/stepfun-ai/Step-3.5-Flash) Step-3.5-Flash HuggingFace event\n[E2](https://huggingface.co/stepfun-ai/Step3-VL-10B) Step3-VL-10B HuggingFace event\n[E3](https://huggingface.co/stepfun-ai/Step-3.7-Flash) Step-3.7-Flash HuggingFace event\n[E4](https://huggingface.co/stepfun-ai/Step-Audio-R1.1) Step-Audio-R1.1 HuggingFace event\n[E5](https://huggingface.co/stepfun-ai/step3) step3 HuggingFace event\n[E6](https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview) GELab-Zero-4B-preview HuggingFace event\n[E7](https://huggingface.co/stepfun-ai/Step-Audio-R1) Step-Audio-R1 HuggingFace event\n[E8](https://huggingface.co/stepfun-ai/Step1X-3D) Step1X-3D HuggingFace event\n[E9](https://huggingface.co/stepfun-ai/NextStep-1-Large) NextStep-1-Large HuggingFace event\n[E10](https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base) Step-3.5-Flash-Base HuggingFace event\n[E11](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2) Step1X-Edit-v1p2 HuggingFace event\n[E12](https://huggingface.co/stepfun-ai/Step3-VL-10B-Base) Step3-VL-10B-Base HuggingFace event\n[E13](https://huggingface.co/stepfun-ai/NextStep-1-Large-Edit) NextStep-1-Large-Edit HuggingFace event\n[E14](https://huggingface.co/stepfun-ai/Step-Audio-AQAA) Step-Audio-AQAA HuggingFace event\n[E15](https://huggingface.co/stepfun-ai/PaCoRe-8B) PaCoRe-8B HuggingFace event\n[E16](https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base-Midtrain) Step-3.5-Flash-Base-Midtrain HuggingFace event\n[E17](https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0) Step-Realtime-CLI v0.1.0 release event\n[E18](https://github.com/stepfun-ai/Step-Video-T2V) Step-Video-T2V repo_new event\n[E19](https://github.com/stepfun-ai/Step-Realtime-CLI) Step-Realtime-CLI repo_new event\n[E20](https://github.com/stepfun-ai/Step-3.7-Flash) Step-3.7-Flash repo_new event\n[E21](https://huggingface.co/stepfun-ai/NextStep-1.1) NextStep-1.1 HuggingFace event\n[E22](https://github.com/stepfun-ai/Step-Audio2) Step-Audio2 repo_new event\n[E23](https://github.com/stepfun-ai/Step1X-Edit) Step1X-Edit repo_new event\n[E24](https://huggingface.co/stepfun-ai/NextStep-1-Large-Pretrain) NextStep-1-Large-Pretrain HuggingFace event\n[E25](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2-preview) Step1X-Edit-v1p2-preview HuggingFace event\n[E26](https://github.com/stepfun-ai/Step-Audio) Step-Audio repo_new event\n[E27](https://huggingface.co/stepfun-ai/NextStep-1-f8ch16-Tokenizer) NextStep-1-f8ch16-Tokenizer HuggingFace event\n[E28](https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain-256px) NextStep-1.1-Pretrain-256px HuggingFace event\n[E29](https://huggingface.co/stepfun-ai/Qwen2.5-32B-DialogueReason) Qwen2.5-32B-DialogueReason HuggingFace event\n[E30](https://github.com/stepfun-ai/PaCoRe) PaCoRe repo_new event\n[E31](https://github.com/stepfun-ai/gelab-zero) gelab-zero repo_new event\n[E32](https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B) StepFun-Formalizer-32B HuggingFace event\n[E33](https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-32B) StepFun-Prover-Preview-32B HuggingFace event\n[E34](https://github.com/stepfun-ai/Step-3.5-Flash) Step-3.5-Flash repo_new event\n[E35](https://huggingface.co/stepfun-ai/Step3-VL-10B-FP8) Step3-VL-10B-FP8 HuggingFace event\n[E36](https://huggingface.co/stepfun-ai/RLVR-8B-0926) RLVR-8B-0926 HuggingFace event\n[E37](https://github.com/stepfun-ai/SteptronOss) SteptronOss repo_new event\n[E38](https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain) NextStep-1.1-Pretrain HuggingFace event\n[E39](https://huggingface.co/stepfun-ai/StepFun-Formalizer-7B) StepFun-Formalizer-7B HuggingFace event\n[E40](https://huggingface.co/stepfun-ai/Step1X-Edit-v1p1-diffusers) Step1X-Edit-v1p1-diffusers HuggingFace event\n[E41](https://github.com/stepfun-ai/Step-Audio-EditX) Step-Audio-EditX repo_new event\n[E42](https://github.com/stepfun-ai/Step1X-3D) Step1X-3D repo_new event\n[E43](https://github.com/stepfun-ai/NextStep-1) NextStep-1 repo_new event\n[E44](https://github.com/stepfun-ai/Step-Audio-R1) Step-Audio-R1 repo_new event\n[E45](https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-7B) StepFun-Prover-Preview-7B HuggingFace event\n[E46](https://github.com/stepfun-ai/StepDeepResearch) StepDeepResearch repo_new event\n[E47](https://github.com/stepfun-ai/Step3) Step3 repo_new event\n[E48](https://github.com/stepfun-ai/Step3-VL-10B) Step3-VL-10B repo_new event\n[E49](https://github.com/stepfun-ai/Step-Video-TI2V) Step-Video-TI2V repo_new event\n[E50](https://github.com/stepfun-ai/StepMesh) StepMesh repo_new event\n[E51](https://github.com/stepfun-ai/Step-Realtime-Console) Step-Realtime-Console repo_new event\n[E52](https://github.com/stepfun-ai/GEBench) GEBench repo_new event\n[E53](https://github.com/stepfun-ai/vllm) vllm fork event\n[E54](https://github.com/stepfun-ai/ComfyUI-StepVideo) ComfyUI-StepVideo repo_new event\n[E55](https://github.com/stepfun-ai/StepFun-Prover-Preview) StepFun-Prover-Preview repo_new event\n[E56](https://github.com/stepfun-ai/StepFun-Formalizer) StepFun-Formalizer repo_new event\n[E57](https://github.com/stepfun-ai/StepAudio-Skills) StepAudio-Skills repo_new event\n[E58](https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark) Step-Audio-Edit-Benchmark repo_new event\n[E59](https://github.com/stepfun-ai/InfiniteHBD-Trace) InfiniteHBD-Trace repo_new event\n[E60](https://github.com/stepfun-ai/llama.cpp) llama.cpp fork event\n[W1](https://static.stepfun.com/blog/step-3.7-flash/) Step 3.7 Flash blog post\n[W2](http://github.com/stepfun-ai/Step-3.7-Flash) Step 3.7 Flash GitHub README\n[W3](https://www.marktechpost.com/2026/05/29/stepfun-releases-step-3-7-flash-a-198b-moe-vision-language-model-for-coding-agents-and-search-workflows/) MarkTechPost coverage of Step 3.7 Flash\n[W4](https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/) NVIDIA Technical Blog on Step 3.7 Flash\n[W5](https://decrypt.co/369013/stepfun-stepaudio-voice-ai-tops-benchmarks) Decrypt article on StepFun Voice AI\n[W6](https://p.rst.im/q/github.com/stepfun-ai/Step-Realtime-CLI) Step-Realtime-CLI GitHub page","generated_at":"2026-06-27T22:28:23.83+00:00","citations":[{"url":"https://huggingface.co/stepfun-ai/Step-3.7-Flash","path":null,"label":"stepfun-ai/Step-3.7-Flash","type":"external"},{"url":"https://static.stepfun.com/blog/step-3.7-flash/","path":null,"label":"static.stepfun.com/blog","type":"external"},{"url":"https://developer.nvidia.com/blog/run-step-3-7-flash-on-nvidia-gpus-with-enterprise-ready-multimodal-ai/","path":null,"label":"developer.nvidia.com/blog","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Realtime-CLI","path":null,"label":"stepfun-ai/Step-Realtime-CLI","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0","path":null,"label":"stepfun-ai/Step-Realtime-CLI","type":"external"},{"url":"http://github.com/stepfun-ai/Step-3.7-Flash","path":null,"label":"stepfun-ai/Step-3.7-Flash","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Realtime-Console","path":null,"label":"stepfun-ai/Step-Realtime-Console","type":"external"},{"url":"https://github.com/stepfun-ai/vllm","path":null,"label":"stepfun-ai/vllm","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio2","path":null,"label":"stepfun-ai/Step-Audio2","type":"external"},{"url":"https://github.com/stepfun-ai/llama.cpp","path":null,"label":"stepfun-ai/llama.cpp","type":"external"},{"url":"https://www.marktechpost.com/2026/05/29/stepfun-releases-step-3-7-flash-a-198b-moe-vision-language-model-for-coding-agents-and-search-workflows/","path":null,"label":"marktechpost.com/2026","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step-3.5-Flash","path":null,"label":"stepfun-ai/Step-3.5-Flash","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base","path":null,"label":"stepfun-ai/Step-3.5-Flash-Base","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step-3.5-Flash-Base-Midtrain","path":null,"label":"stepfun-ai/Step-3.5-Flash-Base-Midtrain","type":"external"},{"url":"https://huggingface.co/stepfun-ai/step3","path":null,"label":"stepfun-ai/step3","type":"external"},{"url":"https://github.com/stepfun-ai/Step3","path":null,"label":"stepfun-ai/Step3","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step3-VL-10B","path":null,"label":"stepfun-ai/Step3-VL-10B","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step3-VL-10B-Base","path":null,"label":"stepfun-ai/Step3-VL-10B-Base","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step3-VL-10B-FP8","path":null,"label":"stepfun-ai/Step3-VL-10B-FP8","type":"external"},{"url":"https://github.com/stepfun-ai/Step3-VL-10B","path":null,"label":"stepfun-ai/Step3-VL-10B","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step-Audio-R1.1","path":null,"label":"stepfun-ai/Step-Audio-R1.1","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step-Audio-R1","path":null,"label":"stepfun-ai/Step-Audio-R1","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio-R1","path":null,"label":"stepfun-ai/Step-Audio-R1","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio-EditX","path":null,"label":"stepfun-ai/Step-Audio-EditX","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio-EditX","path":null,"label":"stepfun-ai/Step-Audio-EditX","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio2","path":null,"label":"stepfun-ai/Step-Audio2","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Video-T2V","path":null,"label":"stepfun-ai/Step-Video-T2V","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Video-TI2V","path":null,"label":"stepfun-ai/Step-Video-TI2V","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Video-T2V","path":null,"label":"stepfun-ai/Step-Video-T2V","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Video-TI2V","path":null,"label":"stepfun-ai/Step-Video-TI2V","type":"external"},{"url":"https://github.com/stepfun-ai/Step1X-Edit","path":null,"label":"stepfun-ai/Step1X-Edit","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2","path":null,"label":"stepfun-ai/Step1X-Edit-v1p2","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step1X-Edit-v1p2-preview","path":null,"label":"stepfun-ai/Step1X-Edit-v1p2-preview","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step1X-Edit-v1p1-diffusers","path":null,"label":"stepfun-ai/Step1X-Edit-v1p1-diffusers","type":"external"},{"url":"https://github.com/stepfun-ai/Step1X-Edit","path":null,"label":"stepfun-ai/Step1X-Edit","type":"external"},{"url":"https://huggingface.co/stepfun-ai/NextStep-1-Large","path":null,"label":"stepfun-ai/NextStep-1-Large","type":"external"},{"url":"https://huggingface.co/stepfun-ai/NextStep-1-Large-Edit","path":null,"label":"stepfun-ai/NextStep-1-Large-Edit","type":"external"},{"url":"https://huggingface.co/stepfun-ai/NextStep-1.1","path":null,"label":"stepfun-ai/NextStep-1.1","type":"external"},{"url":"https://huggingface.co/stepfun-ai/NextStep-1-Large-Pretrain","path":null,"label":"stepfun-ai/NextStep-1-Large-Pretrain","type":"external"},{"url":"https://huggingface.co/stepfun-ai/NextStep-1-f8ch16-Tokenizer","path":null,"label":"stepfun-ai/NextStep-1-f8ch16-Tokenizer","type":"external"},{"url":"https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain-256px","path":null,"label":"stepfun-ai/NextStep-1.1-Pretrain-256px","type":"external"},{"url":"https://huggingface.co/stepfun-ai/NextStep-1.1-Pretrain","path":null,"label":"stepfun-ai/NextStep-1.1-Pretrain","type":"external"},{"url":"https://github.com/stepfun-ai/NextStep-1","path":null,"label":"stepfun-ai/NextStep-1","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step1X-3D","path":null,"label":"stepfun-ai/Step1X-3D","type":"external"},{"url":"https://github.com/stepfun-ai/Step1X-3D","path":null,"label":"stepfun-ai/Step1X-3D","type":"external"},{"url":"https://github.com/stepfun-ai/Step1X-3D","path":null,"label":"stepfun-ai/Step1X-3D","type":"external"},{"url":"https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview","path":null,"label":"stepfun-ai/GELab-Zero-4B-preview","type":"external"},{"url":"https://github.com/stepfun-ai/gelab-zero","path":null,"label":"stepfun-ai/gelab-zero","type":"external"},{"url":"https://github.com/stepfun-ai/gelab-zero","path":null,"label":"stepfun-ai/gelab-zero","type":"external"},{"url":"https://github.com/stepfun-ai/StepDeepResearch","path":null,"label":"stepfun-ai/StepDeepResearch","type":"external"},{"url":"https://github.com/stepfun-ai/StepDeepResearch","path":null,"label":"stepfun-ai/StepDeepResearch","type":"external"},{"url":"https://huggingface.co/stepfun-ai/PaCoRe-8B","path":null,"label":"stepfun-ai/PaCoRe-8B","type":"external"},{"url":"https://github.com/stepfun-ai/PaCoRe","path":null,"label":"stepfun-ai/PaCoRe","type":"external"},{"url":"https://huggingface.co/stepfun-ai/RLVR-8B-0926","path":null,"label":"stepfun-ai/RLVR-8B-0926","type":"external"},{"url":"https://github.com/stepfun-ai/PaCoRe","path":null,"label":"stepfun-ai/PaCoRe","type":"external"},{"url":"https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-32B","path":null,"label":"stepfun-ai/StepFun-Prover-Preview-32B","type":"external"},{"url":"https://huggingface.co/stepfun-ai/StepFun-Prover-Preview-7B","path":null,"label":"stepfun-ai/StepFun-Prover-Preview-7B","type":"external"},{"url":"https://github.com/stepfun-ai/StepFun-Prover-Preview","path":null,"label":"stepfun-ai/StepFun-Prover-Preview","type":"external"},{"url":"https://huggingface.co/stepfun-ai/StepFun-Formalizer-32B","path":null,"label":"stepfun-ai/StepFun-Formalizer-32B","type":"external"},{"url":"https://huggingface.co/stepfun-ai/StepFun-Formalizer-7B","path":null,"label":"stepfun-ai/StepFun-Formalizer-7B","type":"external"},{"url":"https://github.com/stepfun-ai/StepFun-Formalizer","path":null,"label":"stepfun-ai/StepFun-Formalizer","type":"external"},{"url":"https://github.com/stepfun-ai/StepMesh","path":null,"label":"stepfun-ai/StepMesh","type":"external"},{"url":"https://github.com/stepfun-ai/StepMesh","path":null,"label":"stepfun-ai/StepMesh","type":"external"},{"url":"https://github.com/stepfun-ai/SteptronOss","path":null,"label":"stepfun-ai/SteptronOss","type":"external"},{"url":"https://github.com/stepfun-ai/SteptronOss","path":null,"label":"stepfun-ai/SteptronOss","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Realtime-CLI/releases/tag/v0.1.0","path":null,"label":"stepfun-ai/Step-Realtime-CLI","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Realtime-Console","path":null,"label":"stepfun-ai/Step-Realtime-Console","type":"external"},{"url":"https://github.com/stepfun-ai/ComfyUI-StepVideo","path":null,"label":"stepfun-ai/ComfyUI-StepVideo","type":"external"},{"url":"https://github.com/stepfun-ai/ComfyUI-StepVideo","path":null,"label":"stepfun-ai/ComfyUI-StepVideo","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark","path":null,"label":"stepfun-ai/Step-Audio-Edit-Benchmark","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio-Edit-Benchmark","path":null,"label":"stepfun-ai/Step-Audio-Edit-Benchmark","type":"external"},{"url":"https://github.com/stepfun-ai/InfiniteHBD-Trace","path":null,"label":"stepfun-ai/InfiniteHBD-Trace","type":"external"},{"url":"https://github.com/stepfun-ai/InfiniteHBD-Trace","path":null,"label":"stepfun-ai/InfiniteHBD-Trace","type":"external"},{"url":"https://github.com/stepfun-ai/GEBench","path":null,"label":"stepfun-ai/GEBench","type":"external"},{"url":"https://github.com/stepfun-ai/StepAudio-Skills","path":null,"label":"stepfun-ai/StepAudio-Skills","type":"external"},{"url":"https://github.com/stepfun-ai/stepfunApi-audio-sdk","path":null,"label":"stepfun-ai/stepfunApi-audio-sdk","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Qwen2.5-32B-DialogueReason","path":null,"label":"stepfun-ai/Qwen2.5-32B-DialogueReason","type":"external"},{"url":"https://decrypt.co/369013/stepfun-stepaudio-voice-ai-tops-benchmarks","path":null,"label":"decrypt.co/369013","type":"external"},{"url":"https://p.rst.im/q/github.com/stepfun-ai/Step-Realtime-CLI","path":null,"label":"p.rst.im/q","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio","path":null,"label":"stepfun-ai/Step-Audio","type":"external"},{"url":"https://github.com/stepfun-ai/Step-3.5-Flash","path":null,"label":"stepfun-ai/Step-3.5-Flash","type":"external"},{"url":"https://github.com/stepfun-ai/Step-3.7-Flash","path":null,"label":"stepfun-ai/Step-3.7-Flash","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Realtime-Python-Demo","path":null,"label":"stepfun-ai/Step-Realtime-Python-Demo","type":"external"},{"url":"https://github.com/stepfun-ai/Step1X-3D/releases/tag/v.1.0.0","path":null,"label":"stepfun-ai/Step1X-3D","type":"external"},{"url":"https://github.com/stepfun-ai/Step1X-3D/releases/tag/v.1.0.1","path":null,"label":"stepfun-ai/Step1X-3D","type":"external"},{"url":"https://huggingface.co/stepfun-ai/Step-Audio-AQAA","path":null,"label":"stepfun-ai/Step-Audio-AQAA","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Realtime-CLI","path":null,"label":"stepfun-ai/Step-Realtime-CLI","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio","path":null,"label":"stepfun-ai/Step-Audio","type":"external"},{"url":"https://github.com/stepfun-ai/NextStep-1","path":null,"label":"stepfun-ai/NextStep-1","type":"external"},{"url":"https://github.com/stepfun-ai/Step-Audio-R1","path":null,"label":"stepfun-ai/Step-Audio-R1","type":"external"},{"url":"https://github.com/stepfun-ai/Step3","path":null,"label":"stepfun-ai/Step3","type":"external"},{"url":"https://github.com/stepfun-ai/Step3-VL-10B","path":null,"label":"stepfun-ai/Step3-VL-10B","type":"external"},{"url":"https://github.com/stepfun-ai/StepFun-Prover-Preview","path":null,"label":"stepfun-ai/StepFun-Prover-Preview","type":"external"},{"url":"https://github.com/stepfun-ai/StepFun-Formalizer","path":null,"label":"stepfun-ai/StepFun-Formalizer","type":"external"}],"provenance":{"provider":"deepseek","model":"deepseek-v4-pro","workflow":"onlylabs-deepagents-analysis-v3","agent":"deepagents"},"evidence":{"total":94,"pages":28,"events":65,"web":6,"signal_desks":{"forks":2,"repos":27,"hiring":0,"talking":0,"releases":31},"data_radar_lanes":null,"data_radar_matches":null}},"signal_counts":{"total":65,"model_released":30,"release":3,"repo_new":30,"repo_forked":2,"post_published":0,"job_opened":0}}