KI-Modell-Änderungsprotokoll · 2026-05-12

Am 2026-05-12 wurden 124 Ereignisse erfasst.

Modell hinzugefügt (85)

Alibaba (Qwen)New model: Qwen Image
Mistral New model: Mixtral 8x22B
Microsoft New model: Phi-4 15B
nano-gpt New model: Nvidia Nemotron Ultra 253B
nano-gpt New model: Gemma 3 1B IT
nano-gpt New model: DeepSeek TNG R1T2 Chimera
databricks New model: GPT OSS 120B
databricks New model: GPT OSS 20B
databricks New model: Claude Haiku 4.5 (latest)
databricks New model: Claude Sonnet 4.6
databricks New model: Gemini 3 Pro Preview
databricks New model: Gemini 3.1 Pro Preview Custom Tools
databricks New model: GPT-5.4 mini
databricks New model: Claude Opus 4.6
databricks New model: Gemini 2.5 Pro
databricks New model: GPT-5 Nano
databricks New model: GPT-5.4
databricks New model: Gemini 3.1 Flash Lite Preview
databricks New model: GPT-5
databricks New model: Claude Opus 4.1 (latest)
databricks New model: GPT-5.2
databricks New model: GPT-5.5
databricks New model: Gemini 3 Flash Preview
databricks New model: Claude Opus 4.5 (latest)
databricks New model: GPT-5 Mini
databricks New model: Claude Sonnet 4.5
databricks New model: GPT-5.1
databricks New model: Claude Opus 4.7
databricks New model: Gemini 2.5 Flash
databricks New model: GPT-5.4 nano
databricks New model: Claude Sonnet 4.5 (latest)
chutes New model: Qwen3 32B TEE
chutes New model: Qwen3.6 27B TEE
chutes New model: Qwen3 Coder Next TEE
chutes New model: MiMo V2 Flash TEE
chutes New model: gemma 4 31B turbo TEE
chutes New model: DeepSeek TNG R1T2 Chimera TEE
kilo New model: NVIDIA: Llama 3.1 Nemotron 70B Instruct
opencode New model: DeepSeek V4 Flash Free
nebius New model: Llama-3.1-Nemotron-Ultra-253B-v1
helicone New model: DeepSeek TNG R1T2 Chimera
NVIDIA New model: solar-10.7b-instruct
NVIDIA New model: FLUX.2 Klein 4B
NVIDIA New model: FLUX.1-schnell
NVIDIA New model: FLUX.1-Kontext-dev
NVIDIA New model: mistral-nemotron
NVIDIA New model: Mistral Medium 3
NVIDIA New model: Mistral: Mixtral 8x7B Instruct
NVIDIA New model: Magistral Small 2506
NVIDIA New model: Mistral-7B-Instruct-v0.3
NVIDIA New model: mistral-small-4-119b-2603
NVIDIA New model: sarvam-m
NVIDIA New model: usdvalidate
NVIDIA New model: synthetic-video-detector
NVIDIA New model: cosmos-transfer1-7b
NVIDIA New model: rerank-qa-mistral-4b
NVIDIA New model: nv-embedcode-7b-v1
NVIDIA New model: riva-translate-4b-instruct-v1_1
NVIDIA New model: nemotron-voicechat
NVIDIA New model: usdcode
NVIDIA New model: llama-3_2-nemoretriever-300m-embed-v1
NVIDIA New model: llama-3.1-nemotron-safety-guard-8b-v3
NVIDIA New model: nv-embed-v1
NVIDIA New model: gliner-pii
NVIDIA New model: bevformer
NVIDIA New model: nemotron-mini-4b-instruct
NVIDIA New model: cosmos-predict1-5b
NVIDIA New model: cosmos-transfer2.5-2b
NVIDIA New model: nemotron-content-safety-reasoning-4b
NVIDIA New model: Active Speaker Detection
NVIDIA New model: magpie-tts-zeroshot
NVIDIA New model: streampetr
NVIDIA New model: sparsedrive
NVIDIA New model: nemotron-3-content-safety
NVIDIA New model: llama-nemotron-embed-vl-1b-v2
NVIDIA New model: studiovoice
NVIDIA New model: llama-nemotron-rerank-vl-1b-v2
Meta New model: esm2-650m
Meta New model: esmfold
NVIDIA New model: Qwen Image Edit
NVIDIA New model: paligemma
Meta New model: dracarys-llama-3.1-70b-instruct
NVIDIA New model: BGE M3
NVIDIA New model: ByteDance-Seed/Seed-OSS-36B-Instruct
llmgateway New model: Gemma 3 1B IT

Preis geändert (24)

Alibaba (Qwen)Price update: Qwen2.5-Coder 32B Instruct now $0.0272/$0.1087 (was $0.03/$0.11)
Alibaba (Qwen)Price update: Qwen3 Coder Next now $0.12/$0.75 (was $0.07/$0.3)
chutes Price update: DeepHermes 3 Mistral 24B Preview now $0.0245/$0.0978 (was $0.02/$0.1)
chutes Price update: dots.ocr now $0.01/$0.0109 (was $0.01/$0.01)
chutes Price update: GLM 4.7 TEE now $0.39/$1.75 (was $0.4/$1.5)
chutes Price update: GLM 5.1 TEE now $1.05/$3.5 (was $0.95/$3.15)
chutes Price update: GLM 5 TEE now $0.95/$2.55 (was $0.95/$3.15)
chutes Price update: gpt oss 120b TEE now $0.09/$0.36 (was $0.04/$0.18)
chutes Price update: Hermes 4 14B now $0.0136/$0.0543 (was $0.01/$0.05)
chutes Price update: Kimi K2.5 TEE now $0.44/$2 (was $0.6/$3)
chutes Price update: Kimi K2.6 TEE now $0.95/$4 (was $0.44/$2)
chutes Price update: MiniMax M2.5 TEE now $0.15/$1.2 (was $0.3/$1.1)
chutes Price update: Qwen3 235B A22B Instruct 2507 TEE now $0.1/$0.6 (was $0.08/$0.55)
chutes Price update: Qwen3Guard Gen 0.6B now $0.01/$0.0109 (was $0.01/$0.01)
chutes Price update: DeepSeek R1 0528 TEE now $0.45/$2.15 (was $0.4/$1.75)
chutes Price update: DeepSeek V3 0324 TEE now $0.25/$1 (was $0.19/$0.87)
chutes Price update: DeepSeek V3.1 TEE now $0.27/$1 (was $0.2/$0.8)
DeepSeek Price update: DeepSeek R1 Distill Llama 70B now $0.0272/$0.1087 (was $0.03/$0.11)
Google Price update: Gemma 3 27B now $0.0272/$0.1087 (was $0.03/$0.11)
Google Price update: Gemma 3 4B now $0.01/$0.0272 (was $0.01/$0.03)
Meta Price update: Llama 3.2 3b Instruct now $0.01/$0.0136 (was $0.01/$0.01)
Mistral Price update: Mistral Small 3.2 24B Instruct (2506) now $0.1/$0.31 (was $0.06/$0.18)
OpenAI Price update: GPT OSS 20B now $0.03/$0.14 (was $0.02/$0.1)
Z.AI / Zhipu Price update: GLM-4.7-Flash now $0.06/$0.4 (was $0.06/$0.35)

Funktion aktualisiert (7)

chutes Capabilities updated: dots.ocr
chutes Capabilities updated: Kimi K2.5 TEE
chutes Capabilities updated: Qwen3Guard Gen 0.6B
Meta Capabilities updated: Llama 3.2 1B Instruct
Meta Capabilities updated: Llama 3.1 Instruct (8B)
NVIDIA Capabilities updated: Llama 3.3 Nemotron Super 49B v1
NVIDIA Capabilities updated: Llama 3.3 Nemotron Super 49B v1.5

Output-Limit geändert (4)

chutes Max output: Kimi K2.6 TEE updated
Meta Max output: Llama 3.1 Instruct (8B) updated
NVIDIA Max output: Llama 3.3 Nemotron Super 49B v1 updated
NVIDIA Max output: Llama 3.3 Nemotron Super 49B v1.5 updated

Kontext geändert (4)

Meta Context window: Llama 3.2 1B Instruct now 16,384 (was 32,768)
Meta Context window: Llama 3.1 Instruct (8B) now 131,072 (was 128,000)
NVIDIA Context window: Llama 3.3 Nemotron Super 49B v1 now 131,072 (was 128,000)
NVIDIA Context window: Llama 3.3 Nemotron Super 49B v1.5 now 131,072 (was 128,000)

Data is sourced from models.dev and normalized for comparison. Prices and capabilities may change. Always verify critical production decisions with the provider's official documentation.