The Inference Report
The week's announcements reveal a hardware-software arms race in inference optimization and a divergence in how labs are positioning around deployment scale versus research narratives. AMD is shipping four technical posts on inference performance, sparse attention for video, NVFP4 quantization support on MI355, INT3 quantization and tensor parallelism optimization, and automated kernel tuning for DeepSeekV4, each addressing a specific bottleneck in production serving. These are not research papers; they are engineering solutions to problems that matter when models hit real infrastructure. The posts signal AMD's pivot toward making its accelerators compatible with checkpoint formats and optimization techniques that already exist in the ecosystem, removing friction that previously forced costly offline conversion. Google DeepMind, by contrast, announced ATL Saathi, a Gemini-powered tool for Indian robotics educators, which positions the company's inference capability as a distribution channel into emerging markets while building developer familiarity early. Anthropic released two items on Claude's behavior across models and languages, plus a robotics red-teaming effort, work that lives in the research-and-safety narrative space rather than the deployment optimization space. The pattern is clear: infrastructure players are racing to commoditize inference performance and cross-platform compatibility, while frontier labs are spending cycles on values alignment and red-teaming exercises that do not directly improve model capability or reduce serving cost. The money is moving toward whoever can make models cheaper and faster to run on existing hardware; the attention is moving toward whoever can publish the most credible story about model behavior.
Sloane Duvall
A curated reference of models from major AI labs, with open/closed weight status, input modalities, and context window size. American labs tend towards closed weights models and Chinese labs tend toward open weights models.
Amazon- Amazon: Nova 2 LiteTextVisionVideoFiles1M
- Amazon: Nova Premier 1.0TextVision1M
None
Anthropic- Anthropic: Claude Fable 5TextVisionFiles1M
- Anthropic: Claude Haiku 4.5TextVisionFiles200K
- Anthropic: Claude Opus 4.1VisionTextFiles200K
- Anthropic: Claude Opus 4.5FilesVisionText200K
- Anthropic: Claude Opus 4.6TextVisionFiles1M
- Anthropic: Claude Opus 4.7TextVisionFiles1M
- Anthropic: Claude Opus 4.7 (Fast)TextVisionFiles1M
- Anthropic: Claude Opus 4.8TextVisionFiles1M
- Anthropic: Claude Opus 4.8 (Fast)TextVisionFiles1M
- Anthropic: Claude Sonnet 4.5TextVisionFiles1M
- Anthropic: Claude Sonnet 4.6TextVisionFiles1M
- Anthropic: Claude Sonnet 5TextVisionFiles1M
None
Google DeepMind- Google: Gemini 2.5 Flash LiteTextVisionFilesAudioVideo1M
- Google: Gemini 3 Flash PreviewTextVisionFilesAudioVideo1M
- Google: Gemini 3.1 Flash LiteTextVisionVideoFilesAudio1M
- Google: Gemini 3.1 Flash Lite PreviewTextVisionVideoFilesAudio1M
- Google: Gemini 3.1 Pro PreviewAudioFilesVisionTextVideo1M
- Google: Gemini 3.1 Pro Preview Custom ToolsTextAudioVisionVideoFiles1M
- Google: Gemini 3.5 FlashTextVisionVideoFilesAudio1M
- Google: Lyria 3 Clip PreviewTextVision1M
- Google: Lyria 3 Pro PreviewTextVision1M
- Google: Nano Banana (Gemini 2.5 Flash Image)VisionText33K
- Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)VisionText131K
- Google: Nano Banana 2 (Gemini 3.1 Flash Image)VisionText131K
- Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)VisionText66K
- Google: Nano Banana Pro (Gemini 3 Pro Image Preview)VisionText66K
- Google: Nano Banana Pro (Gemini 3 Pro Image)VisionText66K
- Google: Gemma 4 26B A4B VisionTextVideo262K
- Google: Gemma 4 31BVisionTextVideo262K
OpenAI- OpenAI: GPT AudioTextAudio128K
- OpenAI: GPT Audio MiniTextAudio128K
- OpenAI: GPT Chat LatestTextVisionFiles400K
- OpenAI: GPT-5TextVisionFiles400K
- OpenAI: GPT-5 ChatFilesVisionText128K
- OpenAI: GPT-5 CodexTextVision400K
- OpenAI: GPT-5 ImageVisionTextFiles400K
- OpenAI: GPT-5 Image MiniFilesVisionText400K
- OpenAI: GPT-5 MiniTextVisionFiles400K
- OpenAI: GPT-5 NanoTextVisionFiles400K
- OpenAI: GPT-5 ProVisionTextFiles400K
- OpenAI: GPT-5.1VisionTextFiles400K
- OpenAI: GPT-5.1 ChatFilesVisionText128K
- OpenAI: GPT-5.1-CodexTextVision400K
- OpenAI: GPT-5.1-Codex-MaxTextVision400K
- OpenAI: GPT-5.1-Codex-MiniVisionText400K
- OpenAI: GPT-5.2FilesVisionText400K
- OpenAI: GPT-5.2 ChatFilesVisionText128K
- OpenAI: GPT-5.2 ProVisionTextFiles400K
- OpenAI: GPT-5.2-CodexTextVision400K
- OpenAI: GPT-5.3 ChatTextVisionFiles128K
- OpenAI: GPT-5.3-CodexTextVisionFiles400K
- OpenAI: GPT-5.4TextVisionFiles1M
- OpenAI: GPT-5.4 Image 2VisionTextFiles272K
- OpenAI: GPT-5.4 MiniFilesVisionText400K
- OpenAI: GPT-5.4 NanoFilesVisionText400K
- OpenAI: GPT-5.4 ProTextVisionFiles1M
- OpenAI: GPT-5.5FilesVisionText1M
- OpenAI: GPT-5.5 ProFilesVisionText1M
- OpenAI: GPT-5.6 LunaFilesVisionText1M
- OpenAI: GPT-5.6 Luna ProFilesVisionText1M
- OpenAI: GPT-5.6 SolFilesVisionText1M
- OpenAI: GPT-5.6 Sol ProFilesVisionText1M
- OpenAI: GPT-5.6 TerraFilesVisionText1M
- OpenAI: GPT-5.6 Terra ProFilesVisionText1M
- OpenAI: o3 Deep ResearchVisionTextFiles200K
- OpenAI: o4 Mini Deep ResearchFilesVisionText200K
- OpenAI: gpt-oss-120bText131K
- OpenAI: gpt-oss-20bText131K
- OpenAI: gpt-oss-safeguard-20bText131K
xAI- xAI: Grok 4.20TextVisionFiles2M
- xAI: Grok 4.20 Multi-AgentTextVisionFiles2M
- xAI: Grok 4.3TextVisionFiles1M
- xAI: Grok 4.5TextVisionFiles500K
- xAI: Grok Build 0.1TextVisionFiles256K
None
Mistral AI- Mistral: Codestral 2508TextFiles256K
- Mistral: Mistral Large 3 2512TextVisionFiles262K
- Mistral: Mistral Medium 3.1TextVisionFiles131K
- Mistral: Mistral Medium 3.5TextVisionFiles262K
- Mistral: Devstral 2 2512TextFiles262K
- Mistral: Ministral 3 14B 2512TextVision262K
- Mistral: Ministral 3 3B 2512TextVision131K
- Mistral: Ministral 3 8B 2512TextVision262K
- Mistral: Mistral Small 4TextVision262K
- Mistral: Voxtral Small 24B 2507TextAudioFiles32K
AI21 LabsNone
- AI21: Jamba Large 1.7Text256K
Alibaba (Qwen)- Qwen: Qwen Plus 0728Text1M
- Qwen: Qwen Plus 0728 (thinking)Text1M
- Qwen: Qwen3 Coder FlashText1M
- Qwen: Qwen3 Coder PlusText1M
- Qwen: Qwen3 MaxText262K
- Qwen: Qwen3 Max ThinkingText262K
- Qwen: Qwen3.5 Plus 2026-02-15TextVisionVideo1M
- Qwen: Qwen3.5 Plus 2026-04-20TextVisionVideo1M
- Qwen: Qwen3.5-FlashTextVisionVideo1M
- Qwen: Qwen3.6 FlashTextVisionVideo1M
- Qwen: Qwen3.6 Max PreviewText262K
- Qwen: Qwen3.6 PlusTextVisionVideo1M
- Qwen: Qwen3.7 MaxText1M
- Qwen: Qwen3.7 PlusTextVision1M
- Qwen: Qwen3 235B A22B Instruct 2507Text262K
- Qwen: Qwen3 235B A22B Thinking 2507Text262K
- Qwen: Qwen3 30B A3B Instruct 2507Text131K
- Qwen: Qwen3 30B A3B Thinking 2507Text131K
- Qwen: Qwen3 Coder 30B A3B InstructText160K
- Qwen: Qwen3 Coder 480B A35BText1M
- Qwen: Qwen3 Coder NextText262K
- Qwen: Qwen3 Next 80B A3B InstructText262K
- Qwen: Qwen3 Next 80B A3B ThinkingText262K
- Qwen: Qwen3 VL 235B A22B InstructTextVision262K
- Qwen: Qwen3 VL 235B A22B ThinkingTextVision131K
- Qwen: Qwen3 VL 30B A3B InstructTextVision262K
- Qwen: Qwen3 VL 30B A3B ThinkingTextVision131K
- Qwen: Qwen3 VL 32B InstructTextVision262K
- Qwen: Qwen3 VL 8B InstructVisionText256K
- Qwen: Qwen3 VL 8B ThinkingVisionText256K
- Qwen: Qwen3.5 397B A17BTextVisionVideo256K
- Qwen: Qwen3.5-122B-A10BTextVisionVideo262K
- Qwen: Qwen3.5-27BTextVisionVideo262K
- Qwen: Qwen3.5-35B-A3BTextVisionVideo262K
- Qwen: Qwen3.5-9BTextVisionVideo262K
- Qwen: Qwen3.6 27BTextVisionVideo262K
- Qwen: Qwen3.6 35B A3BTextVisionVideo262K
ByteDance- Seed: Seed 1.6VisionTextVideo262K
- Seed: Seed 1.6 FlashVisionTextVideo262K
- Seed: Seed-2.0-LiteTextVisionVideo262K
- Seed: Seed-2.0-MiniTextVisionVideo262K
None
DeepSeekNone
- DeepSeek: DeepSeek V3.1Text164K
- DeepSeek: DeepSeek V3.1 TerminusText164K
- DeepSeek: DeepSeek V3.2Text131K
- DeepSeek: DeepSeek V3.2 ExpText164K
- DeepSeek: DeepSeek V4 FlashText1M
- DeepSeek: DeepSeek V4 ProText1M
MiniMax- MiniMax: MiniMax M2-herText66K
- MiniMax: MiniMax M2Text205K
- MiniMax: MiniMax M2.1Text205K
- MiniMax: MiniMax M2.5Text205K
- MiniMax: MiniMax M2.7Text205K
- MiniMax: MiniMax M3TextVisionVideo1M