Skip to content
Thursday, 19 November 2026
Tenesys AI News
Subscribe

AI Model Updates

What exactly changed in AI models: context, prices, capabilities. Only values stated by sources — never guessed.

ModelsAWS Machine Learning · 19 Oct 2026

AWS Bedrock adds TwelveLabs' Pegasus 1.5 for structured video search

ParameterBeforeNow
Max video length1 hour (Pegasus 1.2)2 hours per request
Output formatSingle text answer per clip queryStructured, timestamped JSON segments matching defined criteria
Input methodsPre-indexed clipsDirect S3 URI or base64 string, no pre-indexing
PromptingText-only queriesText plus reference image (multimodal) prompting
APIsNot specifiedInvokeModel, InvokeModelWithResponseStream (US West/Oregon)
Read more →
ImportantVoice AIAWS Machine Learning · 19 Oct 2026

Deepgram cuts batch transcription costs by porting Nova-3 to AWS Trainium2

ParameterBeforeNow
Hardware targetStandard PyTorch on prior compute optionsAWS Trainium2 (trn2.3xlarge) via Neuron/TorchNeuron
Throughput per dollarBaseline compute options Deepgram previously ran11-70% more audio processed per dollar
Serving runtimePython multiprocessing/multithreaded data-parallel runtimeCustom Rust runtime, 7.5% faster than Python MP baseline at max batch size
Replicas per deviceNot specified8 model replicas (DP8) on one trn2.3xlarge instance
Read more →
ImportantVoice AIOpenAI API Changelog · 16 Oct 2026

OpenAI makes custom voice generation available to every GPT-Live API user

ParameterBeforeNow
Voice creation accessLimited availabilityAvailable to all API customers
Voice creation costNot specifiedFree
Speech generation billingStandard GPT-Live session billingStandard GPT-Live session billing (unchanged)
Built-in voicesExisting set+9 new voices: German (Nuri, Juni, Flitz), French (Brise, Sillage), Hindi (Nira), Japanese (Harema, Shida), Korean (Noeu
Read more →
ImportantModelsHugging Face · 16 Oct 2026

Reflection releases Beam, a 501B open-weight MoE model with 1M-token context

ParameterBeforeNow
Total parametersNot specified501B (MoE)
Active parameters per tokenNot specified23B
Context windowNot specified1,000,000 tokens
LicenseNot specifiedApache 2.0
Tool callingNot specifiedSupported via dedicated grammar and response template
Reasoning effort levelsNot specifiedlow, medium (default), high, xhigh, max
Checkpoint formatsNot specifiedBF16, FP8, NVFP4
Read more →
ImportantModelsHugging Face · 15 Oct 2026

IBM combines transcription, decision-making and chat in a single Granite Switch model

ParameterBeforeNow
Capabilities per endpointSeparate models/endpoints for chat, decisions, and speechChat, decision-making, and ASR transcription unified on one endpoint
VRAM usageNot specifiedApproximately same VRAM as base model alone, despite added capabilities
Decision adaptersNo Decision aLoRA in prior releaseGeneral-purpose Decision aLoRA plus specialized Guardian aLoRA for safety decisions
ASR componentNot specifiedGranite-Speech-5.0-Turbo-CTC embedded, top-ranked on OpenASR Leaderboard
Inference speed (decision throughput, 3B demo)Competitor Jev reported at 10 decisions/secondGranite Switch 3B variant reached 60-100 decisions/second
Read more →
ImportantModelsHugging Face · 15 Oct 2026

Meta releases Muse Spark Open, an 816B-parameter open-weight multimodal MoE model

ParameterBeforeNow
ArchitectureMuse Glimmer: dense feed-forward network per layerMuse Spark Open: Mixture-of-Experts, 256 experts per layer, 8 selected per token
Model sizeNot specified~816B total parameters, ~42B active per token
Context windowNot specifiedUp to 1,000,000 tokens
InputsNot specified (Muse Glimmer: no audio encoder)Text, images, audio, and video
OutputNot specifiedText
Checkpoint formatsNot specifiedBF16 and NVFP4 (4-bit, for NVIDIA Blackwell GPUs)
Tool callingNot specifiedSupported, text-only and multimodal
Read more →
ImportantModelsHugging Face · 15 Oct 2026

Apertus 1.5 adds image/audio understanding, longer context and a thinking mode to the open 8B/70B models

ParameterBeforeNow
Context windowNot specified (Apertus 1.0 baseline)262,144 tokens (4x increase over Apertus 1.0)
Input modalitiesText onlyText, images, and audio (audio understanding experimental); generation remains text-only
ReasoningNo explicit reasoning modeOptional 'thinking mode' that reasons before answering, boosting performance on tasks like AIME
Tool callingStandard tool useImproved, trained on execution-verified tool-use trajectories plus RL/DPO tuning
Training data volumeApertus 1.0 pretraining corpus8B model: 19T total tokens (+4T); 70B model: 17T total tokens (+2T)
LicenseApache 2.0Apache 2.0 (unchanged, fully open weights/data)
Read more →
ImportantCyber SecurityGitHub · 13 Oct 2026

GitHub swaps LLM-based secret scanner for a lightweight ModernBERT classifier

ParameterBeforeNow
ArchitectureRemote LLM-based classification pipelineTwo-stage pipeline: regex candidate generation + fine-tuned ModernBERT classifier (ONNX, 8-bit quantized)
PrecisionBaseline (previous LLM pipeline)13% higher (offline evaluation)
RecallBaseline (previous LLM pipeline)19% higher (offline evaluation)
Owner-weighted acceptanceBaseline (previous LLM pipeline)11.5% higher (online A/B)
Median post-push scan execution timeBaseline (previous LLM pipeline)86% lower
Serving costBaseline (previous LLM pipeline)More than 98% lower
Read more →
ImportantModelsHugging Face · 13 Oct 2026

Liquid AI's new open-weight model runs agentic reasoning locally with just 2B active parameters

ParameterBeforeNow
Context window32K tokens131K tokens
Vocabulary size65K tokens128K tokens
Pre-training tokens~17T tokens~33T tokens
ReasoningNot reasoning-onlyReasoning-only with explicit chain of thought
Active parameters2B (predecessor LFM2-24B-A2B)2B (unchanged)
Tool callingNot specifiedSupported, trained inside agent harnesses (OpenClaw, Hermes Agent, Pi, OpenCode)
Inference speed (Apple M5 Max)Not specified184 tok/s
Inference speed (NVIDIA B300)Not specifiedUp to 21K output tok/s
Read more →
ImportantResearchAmazon Science · 12 Oct 2026

Amazon's 'Priming' Turns Pretrained Transformers Into Efficient Hybrid Models for Under 1% of Training Cost

ParameterBeforeNow
Training cost vs scratch100% (full pretraining)Less than 1% (under 0.5% of token budget)
Throughput1x (source Transformer)Up to 2x at 128,000 tokens context
KV cache sizeFull attention cacheHalf the KV cache at inference (32B GKA hybrid)
Context length testedNot specified128,000 tokens
LicenseNot specifiedApache 2.0
Read more →
ResearchNVIDIA Developer · 12 Oct 2026

NVIDIA and Princeton Cut Decoding Bottlenecks in Hybrid SSM Models

ParameterBeforeNow
State memory trafficFull state read and written every decode stepCheckpoint read each step, written only once per buffer flush — roughly half the traffic
Kernel speed (batch 256, B300)Baseline vLLM kernel1.46x-1.80x faster; full decode step 1.20x-1.47x faster
Speculative decoding throughput (120B-class)Baseline vLLM speculative decoding1.85x-2.10x higher throughput at batch size 512, matched acceptance lengths
Quantization accuracySingle-precision state prone to rounding error accumulation3 of 4 models stay within ~2 points of FP32 accuracy on GPQA-Diamond using dual-precision checkpoint/read
Read more →
ImportantVoice AIMistral AI (Hugging Face) · 8 Oct 2026

Mistral releases open-source real-time Arabic speech-to-text model Voxtral Mini 4B Realtime Arabic

ParameterBeforeNow
ParametersNot specified~4.4 billion
PrecisionNot specifiedBF16
Audio inputNot specified16 kHz streaming audio
Transcription delayNot specifiedConfigurable, 480 ms tested
Accuracy (CER)Voxtral Transcribe Arabic: 7.91%8.82% average across seven Arabic benchmarks
LicenseNot specifiedApache 2.0
Read more →
ImportantModelsGoogle Developers · 8 Oct 2026

Google Open-Sources ML Drift, a New GPU Engine for On-Device AI Inference

ParameterBeforeNow
Tensor dimensionality4D tensors only (TFLite GPU delegate)5D tensor support out-of-the-box
Platform coverageOpenGL ES, OpenCL, Metal backends separately maintainedUnified shader model via tensor virtualization across OpenGL ES, OpenCL, Metal, WebGPU, plus native Windows/Linux via Da
Memory overheadBaseline (legacy GPU delegate)Up to 12% lower memory overhead in Gemma benchmarks
LLM inferenceNo stage-aware optimizationStage-aware kernel switching for prefill vs decode, custom KV cache layout, in-kernel activation quantization
Read more →
ImportantModelsClaude Platform Release Notes · 7 Oct 2026

Anthropic releases Claude Haiku 5.5 with 1M token context and adaptive thinking

ParameterBeforeNow
Context windowNot specified (Claude Haiku 4.5)1,000,000 tokens
Max output tokensNot specified128,000 tokens
Reasoning / thinking controlManual extended thinking via budget_tokens parameterAdaptive thinking on by default, controlled via effort parameter
AvailabilityClaude Haiku 4.5 on Claude API, Bedrock, AWS, Google Cloud, Microsoft FoundryClaude Haiku 5.5 on Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud, Microsoft Foundry
Read more →
Voice AIHugging Face · 7 Oct 2026

TII releases Falcon-ASR, a 1.6B speech recognition model focused on Arabic and Emirati dialect

ParameterBeforeNow
ParametersNot specified1.6B
LanguagesNot specifiedArabic (incl. Emirati dialect), English, French, Spanish, Portuguese
Arabic avg WER23.17% (best published)20.92%
Emirati WER (internal eval)26.80% (next best, Qwen3-Omni)22.73%
English avg WER (7 public sets)Not specified5.74%
TimestampsNot specifiedWord-level timestamps supported
API accessNot availablePlanned (not yet available)
Read more →
ImportantVoice AIDeepgram · 7 Oct 2026

Deepgram Extends Medical Speech Recognition to 10 Languages

ParameterBeforeNow
Languages supportedEnglish only (Nova-3 Medical)10 languages: Dutch, English, French, German, Hindi, Italian, Japanese, Portuguese, Russian, Spanish
Batch Word Error Rate (vs general multilingual model)6.62%4.96%
Streaming Word Error Rate (vs general multilingual model)8.32%5.86%
Batch Entity Error Rate14.46%9.11%
Streaming Entity Error Rate16.31%9.78%
Code switchingNot specifiedAutomatic code switching between supported languages in both batch and streaming
Read more →
ImportantVoice AIAssemblyAI · 6 Oct 2026

AssemblyAI adds Qwen3.5 4B to LLM Gateway for fast voice-transcript rewriting

ParameterBeforeNow
Context windowNot specified32k
Input priceNot specified$0.10 per million tokens
Output priceNot specified$0.50 per million tokens
Latency (voice rewrite avg)GPT-4.1: 1,138 ms612 ms
Cost per hour of audioGPT-4.1: $0.1546/hr$0.0092/hr
Tool callingNot specifiedNot supported
Supported parametersNot specifiedmax_tokens, temperature, stream only
Read more →