Log in

Changelog

Track the latest AI model benchmarks, performance evaluations, and platform updates.

Changelog begins 24 Dec 2025 for image, video, and speech entries

09 Apr 2026

🔔 New article published
A new look for Artificial Analysis

A new look for Artificial Analysis

Same mission, new look

🔔 New article published
Muse Spark: Meta is back in the AI race

Muse Spark: Meta is back in the AI race

Benchmarks and Analysis of Muse Spark

🔔 New article published
Gemma 4: Google on the American open weights frontier

Gemma 4: Google on the American open weights frontier

Benchmarks and Analysis of Gemma 4 models

🔔 New article published
Sarvam 105B and Sarvam 30B: India enters the open-weights race

Sarvam 105B and Sarvam 30B: India enters the open-weights race

Benchmarks and Analysis of Sarvam 105B and Sarvam 30B

New language model evaluation results available
GLM-5.1 (Non-reasoning)

GLM-5.1 (Non-reasoning)

Intelligence Index: 44 | AA-Omniscience: -21

GLM-5.1 (Non-reasoning) on SiliconFlow

GLM-5.1 (Non-reasoning) on SiliconFlow

Performance benchmark now available
GLM-5.1 (Non-reasoning) on Novita

GLM-5.1 (Non-reasoning) on Novita

Performance benchmark now available
GLM-5.1 (Non-reasoning) on Parasail

GLM-5.1 (Non-reasoning) on Parasail

Performance benchmark now available
GLM-5.1 (Non-reasoning) on DeepInfra

GLM-5.1 (Non-reasoning) on DeepInfra

Performance benchmark now available
GLM-5.1 (Non-reasoning) on FriendliAI

GLM-5.1 (Non-reasoning) on FriendliAI

Performance benchmark now available

08 Apr 2026

New language model evaluation results available
Muse Spark

Muse Spark

Intelligence Index: 52 | AA-Omniscience: 4

New language model evaluation results available
Trinity Large Thinking

Trinity Large Thinking

Intelligence Index: 32 | AA-Omniscience: -44

New language model evaluation results available
Grok 4.20 0309 v2 (Reasoning)

Grok 4.20 0309 v2 (Reasoning)

Intelligence Index: 49 | AA-Omniscience: 15

New language model evaluation results available
Grok 4.20 0309 v2 (Non-reasoning)

Grok 4.20 0309 v2 (Non-reasoning)

Intelligence Index: 29 | AA-Omniscience: -45

New language model evaluation results available
Qwen3.6 Plus

Qwen3.6 Plus

Intelligence Index: 50 | AA-Omniscience: 3

Trinity Large Thinking on Arcee AI

Trinity Large Thinking on Arcee AI

Performance benchmark now available
Trinity Large Thinking on Parasail

Trinity Large Thinking on Parasail

Performance benchmark now available
Grok 4.20 0309 v2 (Reasoning) on xAI

Grok 4.20 0309 v2 (Reasoning) on xAI

Performance benchmark now available
Grok 4.20 0309 v2 (Non-reasoning) on xAI

Grok 4.20 0309 v2 (Non-reasoning) on xAI

Performance benchmark now available
Qwen3.6 Plus on Alibaba Cloud

Qwen3.6 Plus on Alibaba Cloud

Performance benchmark now available

07 Apr 2026

New language model evaluation results available
MiMo-V2-Omni-0327

MiMo-V2-Omni-0327

Intelligence Index: 45 | AA-Omniscience: -14

New language model evaluation results available
GLM-5.1 (Reasoning)

GLM-5.1 (Reasoning)

Intelligence Index: 51 | AA-Omniscience: 2

New language model evaluation results available
Qwen3.5 Omni Flash

Qwen3.5 Omni Flash

Intelligence Index: 26 | AA-Omniscience: -66

GLM-5.1 (Reasoning) on SiliconFlow

GLM-5.1 (Reasoning) on SiliconFlow

Performance benchmark now available
GLM-5.1 (Reasoning) on FriendliAI

GLM-5.1 (Reasoning) on FriendliAI

Performance benchmark now available
GLM-5.1 (Reasoning) on DeepInfra

GLM-5.1 (Reasoning) on DeepInfra

Performance benchmark now available
GLM-5.1 (Reasoning) on Fireworks

GLM-5.1 (Reasoning) on Fireworks

Performance benchmark now available
GLM-5.1 (Reasoning) on GMI

GLM-5.1 (Reasoning) on GMI

Performance benchmark now available
GLM-5.1 (Reasoning) on Novita

GLM-5.1 (Reasoning) on Novita

Performance benchmark now available
GLM-5.1 (Reasoning) on Parasail

GLM-5.1 (Reasoning) on Parasail

Performance benchmark now available
Qwen3.5 Omni Flash on Alibaba Cloud

Qwen3.5 Omni Flash on Alibaba Cloud

Performance benchmark now available
Nova 2.0 Lite (high) on Amazon Bedrock

Nova 2.0 Lite (high) on Amazon Bedrock

Performance benchmark now available
Nova 2.0 Omni (medium) on Amazon Bedrock

Nova 2.0 Omni (medium) on Amazon Bedrock

Performance benchmark now available
Nova 2.0 Lite (high) on Amazon Bedrock

Nova 2.0 Lite (high) on Amazon Bedrock

Performance benchmark now available

06 Apr 2026

03 Apr 2026

02 Apr 2026

🔔 New article published
MAI-Transcribe-1: Everything you need to know

MAI-Transcribe-1: Everything you need to know

Benchmarks and Analysis of MAI-Transcribe-1

🔔 New article published
MiniMax M2.7: Everything you need to know

MiniMax M2.7: Everything you need to know

Benchmarks and Analysis of MiniMax M2.7

🔔 New article published
MiMo-V2-Pro: Everything you need to know

MiMo-V2-Pro: Everything you need to know

Benchmarks and Analysis of MiMo-V2-Pro

New language model evaluation results available
Qwen3.5 Omni Plus

Qwen3.5 Omni Plus

Intelligence Index: 39 | AA-Omniscience: -12

New language model evaluation results available
Gemma 4 26B A4B (Reasoning)

Gemma 4 26B A4B (Reasoning)

Intelligence Index: 31 | AA-Omniscience: -48

New language model evaluation results available
Gemma 4 31B (Reasoning)

Gemma 4 31B (Reasoning)

Intelligence Index: 39 | AA-Omniscience: -45

Qwen3.5 Omni Plus on Alibaba Cloud

Qwen3.5 Omni Plus on Alibaba Cloud

Performance benchmark now available
New Speech to Text model evaluation results available
MAI-Transcribe-1, Azure

MAI-Transcribe-1, Azure

01 Apr 2026

30 Mar 2026

New language model evaluation results available
Nemotron Cascade 2 30B A3B

Nemotron Cascade 2 30B A3B

Intelligence Index: 28 | AA-Omniscience: -55

New language model evaluation results available
KAT Coder Pro V2

KAT Coder Pro V2

Intelligence Index: 44 | AA-Omniscience: -22

Qwen3.5 0.8B (Non-reasoning) on DeepInfra

Qwen3.5 0.8B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 2B (Non-reasoning) on DeepInfra

Qwen3.5 2B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 4B (Non-reasoning) on DeepInfra

Qwen3.5 4B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 27B (Non-reasoning) on DeepInfra

Qwen3.5 27B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 9B (Non-reasoning) on DeepInfra

Qwen3.5 9B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 35B A3B (Non-reasoning) on DeepInfra

Qwen3.5 35B A3B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 122B A10B (Non-reasoning) on DeepInfra

Qwen3.5 122B A10B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 397B A17B (Non-reasoning) on DeepInfra

Qwen3.5 397B A17B (Non-reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 0.8B (Reasoning) on DeepInfra

Qwen3.5 0.8B (Reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 2B (Reasoning) on DeepInfra

Qwen3.5 2B (Reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 4B (Reasoning) on DeepInfra

Qwen3.5 4B (Reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 27B (Reasoning) on DeepInfra

Qwen3.5 27B (Reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 9B (Reasoning) on DeepInfra

Qwen3.5 9B (Reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 35B A3B (Reasoning) on DeepInfra

Qwen3.5 35B A3B (Reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 397B A17B (Reasoning) on DeepInfra

Qwen3.5 397B A17B (Reasoning) on DeepInfra

Performance benchmark now available
Qwen3.5 122B A10B (Reasoning) on DeepInfra

Qwen3.5 122B A10B (Reasoning) on DeepInfra

Performance benchmark now available
KAT Coder Pro V2 on StreamLake

KAT Coder Pro V2 on StreamLake

Performance benchmark now available

26 Mar 2026

25 Mar 2026

24 Mar 2026

23 Mar 2026

21 Mar 2026

20 Mar 2026

19 Mar 2026

New language model evaluation results available
GPT-5.4 nano (Non-Reasoning)

GPT-5.4 nano (Non-Reasoning)

Intelligence Index: 24 | AA-Omniscience: -39

New language model evaluation results available
GPT-5.4 nano (medium)

GPT-5.4 nano (medium)

Intelligence Index: 38 | AA-Omniscience: -19

GPT-5.4 mini (medium) on OpenAI

GPT-5.4 mini (medium) on OpenAI

Performance benchmark now available
GPT-5.4 (xhigh) on Microsoft Azure

GPT-5.4 (xhigh) on Microsoft Azure

Performance benchmark now available
GPT-5.4 nano (medium) on OpenAI

GPT-5.4 nano (medium) on OpenAI

Performance benchmark now available
GPT-5.4 nano (Non-Reasoning) on OpenAI

GPT-5.4 nano (Non-Reasoning) on OpenAI

Performance benchmark now available
GPT-5.4 mini (xhigh) on OpenAI

GPT-5.4 mini (xhigh) on OpenAI

Performance benchmark now available
GPT-5.4 mini (Non-Reasoning) on Microsoft Azure

GPT-5.4 mini (Non-Reasoning) on Microsoft Azure

Performance benchmark now available
GPT-5.4 mini (medium) on Microsoft Azure

GPT-5.4 mini (medium) on Microsoft Azure

Performance benchmark now available
GPT-5.4 mini (xhigh) on Microsoft Azure

GPT-5.4 mini (xhigh) on Microsoft Azure

Performance benchmark now available
GPT-5.4 mini (Non-Reasoning) on OpenAI

GPT-5.4 mini (Non-Reasoning) on OpenAI

Performance benchmark now available
GPT-5.4 nano (xhigh) on OpenAI

GPT-5.4 nano (xhigh) on OpenAI

Performance benchmark now available

18 Mar 2026

New language model evaluation results available
GPT-5.4 mini (Non-Reasoning)

GPT-5.4 mini (Non-Reasoning)

Intelligence Index: 23 | AA-Omniscience: -46

New language model evaluation results available
GPT-5.4 mini (medium)

GPT-5.4 mini (medium)

Intelligence Index: 38 | AA-Omniscience: -20

New language model evaluation results available
GPT-5.4 mini (xhigh)

GPT-5.4 mini (xhigh)

Intelligence Index: 49 | AA-Omniscience: -19

New language model evaluation results available
GPT-5.4 nano (xhigh)

GPT-5.4 nano (xhigh)

Intelligence Index: 44 | AA-Omniscience: -30

New language model evaluation results available
Sarvam 105B (high)

Sarvam 105B (high)

Intelligence Index: 18 | AA-Omniscience: -60

New language model evaluation results available
Sarvam 30B (high)

Sarvam 30B (high)

Intelligence Index: 12 | AA-Omniscience: -72

New language model evaluation results available
GPT-5.4 (Non-reasoning)

GPT-5.4 (Non-reasoning)

Intelligence Index: 35 | AA-Omniscience: -16

New language model evaluation results available
MiniMax-M2.7

MiniMax-M2.7

Intelligence Index: 50 | AA-Omniscience: 1

New language model evaluation results available
Mistral Small 4 (Non-reasoning)

Mistral Small 4 (Non-reasoning)

Intelligence Index: 19 | AA-Omniscience: -48

New language model evaluation results available
Mistral Small 4 (Reasoning)

Mistral Small 4 (Reasoning)

Intelligence Index: 28 | AA-Omniscience: -30

Sarvam 30B (high) on Sarvam

Sarvam 30B (high) on Sarvam

Performance benchmark now available
Sarvam 105B (high) on Sarvam

Sarvam 105B (high) on Sarvam

Performance benchmark now available
MiniMax-M2.5 on Nebius

MiniMax-M2.5 on Nebius

Performance benchmark now available
MiniMax-M2.7 on MiniMax

MiniMax-M2.7 on MiniMax

Performance benchmark now available
MiniMax-M2.7 on Novita

MiniMax-M2.7 on Novita

Performance benchmark now available
Mistral Small 4 (Non-reasoning) on Mistral

Mistral Small 4 (Non-reasoning) on Mistral

Performance benchmark now available
Mistral Small 4 (Reasoning) on Mistral

Mistral Small 4 (Reasoning) on Mistral

Performance benchmark now available

17 Mar 2026

16 Mar 2026

13 Mar 2026

12 Mar 2026

11 Mar 2026

🔔 New article published
NVIDIA Nemotron 3 Super: The new leader in open, efficient intelligence

NVIDIA Nemotron 3 Super: The new leader in open, efficient intelligence

Independent analysis and benchmarks of NVIDIA Nemotron 3 Super

New language model evaluation results available
Grok 4.20 0309 (Non-reasoning)

Grok 4.20 0309 (Non-reasoning)

Intelligence Index: 30 | AA-Omniscience: -43

New language model evaluation results available
Grok 4.20 0309 (Reasoning)

Grok 4.20 0309 (Reasoning)

Intelligence Index: 48 | AA-Omniscience: 13

New language model evaluation results available
Qwen3.5 0.8B (Non-reasoning)

Qwen3.5 0.8B (Non-reasoning)

Intelligence Index: 10 | AA-Omniscience: -89

New language model evaluation results available
NVIDIA Nemotron 3 Super 120B A12B (Reasoning)

NVIDIA Nemotron 3 Super 120B A12B (Reasoning)

Intelligence Index: 36 | AA-Omniscience: -42

New language model evaluation results available
LongCat Flash Lite

LongCat Flash Lite

Intelligence Index: 24 | AA-Omniscience: -70

Grok 4.20 0309 (Reasoning) on xAI

Grok 4.20 0309 (Reasoning) on xAI

Performance benchmark now available
Grok 4.20 0309 (Non-reasoning) on xAI

Grok 4.20 0309 (Non-reasoning) on xAI

Performance benchmark now available
NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on Weights & Biases

NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on Weights & Biases

Performance benchmark now available
NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on Nebius

NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on Nebius

Performance benchmark now available
NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on Baseten

NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on Baseten

Performance benchmark now available
NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on DeepInfra

NVIDIA Nemotron 3 Super 120B A12B (Reasoning) on DeepInfra

Performance benchmark now available
LongCat Flash Lite on LongCat

LongCat Flash Lite on LongCat

Performance benchmark now available

10 Mar 2026

🔔 New article published
Overview of Google I/O Benchmarking Results

Overview of Google I/O Benchmarking Results

Independent benchmarking results covering launches across LLMs, image, video, music and more

Qwen3.5 397B A17B (Reasoning) on Eigen AI

Qwen3.5 397B A17B (Reasoning) on Eigen AI

Performance benchmark now available
MiniMax-M2.5 on Eigen AI

MiniMax-M2.5 on Eigen AI

Performance benchmark now available
Qwen3 8B (Non-reasoning) on Eigen AI

Qwen3 8B (Non-reasoning) on Eigen AI

Performance benchmark now available
Qwen3 30B A3B (Reasoning) on Eigen AI

Qwen3 30B A3B (Reasoning) on Eigen AI

Performance benchmark now available
Qwen3 VL 30B A3B (Reasoning) on Eigen AI

Qwen3 VL 30B A3B (Reasoning) on Eigen AI

Performance benchmark now available
Qwen3 VL 30B A3B Instruct on Eigen AI

Qwen3 VL 30B A3B Instruct on Eigen AI

Performance benchmark now available
Qwen3 8B (Reasoning) on Eigen AI

Qwen3 8B (Reasoning) on Eigen AI

Performance benchmark now available
Qwen3 30B A3B (Non-reasoning) on Eigen AI

Qwen3 30B A3B (Non-reasoning) on Eigen AI

Performance benchmark now available
Qwen3.5 397B A17B (Non-reasoning) on Eigen AI

Qwen3.5 397B A17B (Non-reasoning) on Eigen AI

Performance benchmark now available
Kimi K2.5 (Reasoning) on Clarifai

Kimi K2.5 (Reasoning) on Clarifai

Performance benchmark now available
New model in Text to Video Leaderboard
LTX-2.3 Fast

LTX-2.3 Fast

Text to Video ELO: 1123

New model in Text to Video Leaderboard
LTX-2.3 Pro

LTX-2.3 Pro

Text to Video ELO: 1105

New model in Image to Video Leaderboard
LTX-2.3 Fast

LTX-2.3 Fast

Image to Video ELO: 1169

New model in Image to Video Leaderboard
LTX-2.3 Pro

LTX-2.3 Pro

Image to Video ELO: 1159

09 Mar 2026

07 Mar 2026

06 Mar 2026

New language model evaluation results available
GPT-5.4 Pro (xhigh)

GPT-5.4 Pro (xhigh)

GPT-5.4 (Non-reasoning) on OpenAI

GPT-5.4 (Non-reasoning) on OpenAI

Performance benchmark now available
Kimi K2.5 (Reasoning) on DeepInfra

Kimi K2.5 (Reasoning) on DeepInfra

Performance benchmark now available
Qwen2.5 Instruct 72B on SiliconFlow

Qwen2.5 Instruct 72B on SiliconFlow

Performance benchmark now available
MiniMax-M2.5 on FriendliAI

MiniMax-M2.5 on FriendliAI

Performance benchmark now available
MiniMax-M2.1 on FriendliAI

MiniMax-M2.1 on FriendliAI

Performance benchmark now available
GLM-5 (Reasoning) on FriendliAI

GLM-5 (Reasoning) on FriendliAI

Performance benchmark now available
Qwen3 235B A22B 2507 Instruct on FriendliAI

Qwen3 235B A22B 2507 Instruct on FriendliAI

Performance benchmark now available
MiniMax-M2.5 on Weights & Biases

MiniMax-M2.5 on Weights & Biases

Performance benchmark now available
Kimi K2.5 (Reasoning) on Weights & Biases

Kimi K2.5 (Reasoning) on Weights & Biases

Performance benchmark now available
GLM-5 (Reasoning) on Weights & Biases

GLM-5 (Reasoning) on Weights & Biases

Performance benchmark now available
GPT-5.4 Pro (xhigh) on OpenAI

GPT-5.4 Pro (xhigh) on OpenAI

Performance benchmark now available
Step 3.5 Flash on SiliconFlow

Step 3.5 Flash on SiliconFlow

Performance benchmark now available

05 Mar 2026

🔔 New article published
Qwen3.5 small models: Everything you need to know

Qwen3.5 small models: Everything you need to know

Independent analysis and benchmarks of Qwen 3.5 small models

GPT-5.4 (xhigh) on OpenAI

GPT-5.4 (xhigh) on OpenAI

Performance benchmark now available
GLM-5 (Reasoning) on Baseten

GLM-5 (Reasoning) on Baseten

Performance benchmark now available
GLM-5 (Non-reasoning) on Baseten

GLM-5 (Non-reasoning) on Baseten

Performance benchmark now available
Llama 3.1 Instruct 8B on Weights & Biases

Llama 3.1 Instruct 8B on Weights & Biases

Performance benchmark now available
Qwen3 235B A22B 2507 Instruct on Weights & Biases

Qwen3 235B A22B 2507 Instruct on Weights & Biases

Performance benchmark now available
Qwen3 Coder 480B A35B Instruct on Weights & Biases

Qwen3 Coder 480B A35B Instruct on Weights & Biases

Performance benchmark now available
DeepSeek V3.1 (Non-reasoning) on Weights & Biases

DeepSeek V3.1 (Non-reasoning) on Weights & Biases

Performance benchmark now available
gpt-oss-20B (high) on Weights & Biases

gpt-oss-20B (high) on Weights & Biases

Performance benchmark now available
gpt-oss-120B (high) on Weights & Biases

gpt-oss-120B (high) on Weights & Biases

Performance benchmark now available
Qwen3 235B A22B 2507 (Reasoning) on Weights & Biases

Qwen3 235B A22B 2507 (Reasoning) on Weights & Biases

Performance benchmark now available
Llama 3.3 Instruct 70B on Weights & Biases

Llama 3.3 Instruct 70B on Weights & Biases

Performance benchmark now available
Llama 4 Scout on Weights & Biases

Llama 4 Scout on Weights & Biases

Performance benchmark now available
Phi-4 Mini Instruct on Weights & Biases

Phi-4 Mini Instruct on Weights & Biases

Performance benchmark now available
gpt-oss-120B (low) on Weights & Biases

gpt-oss-120B (low) on Weights & Biases

Performance benchmark now available
gpt-oss-20B (low) on Weights & Biases

gpt-oss-20B (low) on Weights & Biases

Performance benchmark now available
GLM-4.7 (Non-reasoning) on Amazon Bedrock

GLM-4.7 (Non-reasoning) on Amazon Bedrock

Performance benchmark now available
GLM-4.7-Flash (Non-reasoning) on Amazon Bedrock

GLM-4.7-Flash (Non-reasoning) on Amazon Bedrock

Performance benchmark now available
DeepSeek V3.2 (Reasoning) on Amazon Bedrock

DeepSeek V3.2 (Reasoning) on Amazon Bedrock

Performance benchmark now available
DeepSeek V3.2 (Non-reasoning) on Amazon Bedrock

DeepSeek V3.2 (Non-reasoning) on Amazon Bedrock

Performance benchmark now available
GLM-4.7 (Reasoning) on Amazon Bedrock

GLM-4.7 (Reasoning) on Amazon Bedrock

Performance benchmark now available
GLM-4.7-Flash (Reasoning) on Amazon Bedrock

GLM-4.7-Flash (Reasoning) on Amazon Bedrock

Performance benchmark now available

04 Mar 2026

New language model evaluation results available
Qwen3.5 2B (Reasoning)

Qwen3.5 2B (Reasoning)

Intelligence Index: 16 | AA-Omniscience: -47

New language model evaluation results available
Qwen3.5 0.8B (Reasoning)

Qwen3.5 0.8B (Reasoning)

Intelligence Index: 11 | AA-Omniscience: -35

New language model evaluation results available
Qwen3.5 4B (Reasoning)

Qwen3.5 4B (Reasoning)

Intelligence Index: 27 | AA-Omniscience: -57

New language model evaluation results available
Qwen3.5 9B (Reasoning)

Qwen3.5 9B (Reasoning)

Intelligence Index: 32 | AA-Omniscience: -53

Kimi K2.5 (Reasoning) on Amazon Bedrock

Kimi K2.5 (Reasoning) on Amazon Bedrock

Performance benchmark now available
GLM-4.7 (Non-reasoning) on Nebius

GLM-4.7 (Non-reasoning) on Nebius

Performance benchmark now available
Kimi K2.5 (Non-reasoning) on Nebius

Kimi K2.5 (Non-reasoning) on Nebius

Performance benchmark now available
Qwen3 235B A22B 2507 (Reasoning) on Nebius

Qwen3 235B A22B 2507 (Reasoning) on Nebius

Performance benchmark now available
Gemma 3 27B Instruct on Nebius

Gemma 3 27B Instruct on Nebius

Performance benchmark now available

03 Mar 2026

02 Mar 2026

New language model evaluation results available
Qwen3.5 122B A10B (Non-reasoning)

Qwen3.5 122B A10B (Non-reasoning)

Intelligence Index: 36 | AA-Omniscience: -54

New language model evaluation results available
Qwen3.5 27B (Non-reasoning)

Qwen3.5 27B (Non-reasoning)

Intelligence Index: 37 | AA-Omniscience: -47

New language model evaluation results available
Qwen3.5 35B A3B (Non-reasoning)

Qwen3.5 35B A3B (Non-reasoning)

Intelligence Index: 31 | AA-Omniscience: -62

Qwen3.5 27B (Non-reasoning) on Alibaba Cloud

Qwen3.5 27B (Non-reasoning) on Alibaba Cloud

Performance benchmark now available
Qwen3.5 35B A3B (Non-reasoning) on Alibaba Cloud

Qwen3.5 35B A3B (Non-reasoning) on Alibaba Cloud

Performance benchmark now available
Qwen3.5 122B A10B (Non-reasoning) on Alibaba Cloud

Qwen3.5 122B A10B (Non-reasoning) on Alibaba Cloud

Performance benchmark now available
Kimi K2.5 (Reasoning) on Lightning AI

Kimi K2.5 (Reasoning) on Lightning AI

Performance benchmark now available
New model in Text to Video Leaderboard
P-Video

P-Video

Text to Video ELO: 1068

New model in Image to Video Leaderboard
P-Video

P-Video

Image to Video ELO: 1136

01 Mar 2026

Artificial Analysis performance benchmarking methodology
Language model performance benchmarking methodology update (v2.2.0): Expanded prompt set and new default prompt length

Language model performance benchmarking methodology update (v2.2.0): Expanded prompt set and new default prompt length

What's changed

  • Refreshed prompt bank: We expanded and diversified prompts across more task types to better reflect real-world usage. This is important because techniques like speculative decoding mean that we see variation in output speeds depending on the type of output.
  • New default workload: Site-default speed now uses 10k input tokens (was 1k)and we have deprecated performance measurements for 100 token input workloads. 1k, 10k, and 100k remain available via Prompt Options, and the Output Speed by Input Token Count chart is the quickest way to compare workload shapes.

Server location, testing frequency, and median reporting are unchanged.

27 Feb 2026

26 Feb 2026

25 Feb 2026

24 Feb 2026

23 Feb 2026

20 Feb 2026

19 Feb 2026

18 Feb 2026

🔔 New article published
Sonnet 4.6 - Everything you need to know

Sonnet 4.6 - Everything you need to know

Independent analysis and benchmarks of Sonnet 4.6

🔔 New article published
AA-WER v2.0: Speech to Text Accuracy Benchmark

AA-WER v2.0: Speech to Text Accuracy Benchmark

Introducing AA-AgentTalk, a new proprietary evaluation dataset focused on voice agent speech, alongside cleaned ground truth transcripts for VoxPopuli and Earnings22.

New language model evaluation results available
Tri-21B-think Preview

Tri-21B-think Preview

Intelligence Index: 20 | AA-Omniscience: -55

New language model evaluation results available
Tri-21B-Think

Tri-21B-Think

Intelligence Index: 19 | AA-Omniscience: -63

New language model evaluation results available
Qwen3.5 397B A17B (Non-reasoning)

Qwen3.5 397B A17B (Non-reasoning)

Intelligence Index: 40 | AA-Omniscience: -36

Claude Sonnet 4.6 (Non-reasoning, Low Effort) on Anthropic

Claude Sonnet 4.6 (Non-reasoning, Low Effort) on Anthropic

Performance benchmark now available
Qwen3.5 397B A17B (Non-reasoning) on Novita

Qwen3.5 397B A17B (Non-reasoning) on Novita

Performance benchmark now available
Qwen3.5 397B A17B (Reasoning) on Parasail

Qwen3.5 397B A17B (Reasoning) on Parasail

Performance benchmark now available
MiniMax-M2.5 on Parasail

MiniMax-M2.5 on Parasail

Performance benchmark now available
MiniMax-M2.5 on Together.ai

MiniMax-M2.5 on Together.ai

Performance benchmark now available
GLM-5 (Reasoning) on Together.ai

GLM-5 (Reasoning) on Together.ai

Performance benchmark now available
MiniMax-M2.5 on SiliconFlow

MiniMax-M2.5 on SiliconFlow

Performance benchmark now available
MiniMax-M2.5 on SambaNova

MiniMax-M2.5 on SambaNova

Performance benchmark now available
GLM-5 (Reasoning) on Google

GLM-5 (Reasoning) on Google

Performance benchmark now available
Claude Sonnet 4.6 (Non-reasoning, High Effort) on Amazon Bedrock

Claude Sonnet 4.6 (Non-reasoning, High Effort) on Amazon Bedrock

Performance benchmark now available
Claude Sonnet 4.6 (Non-reasoning, High Effort) on Microsoft Azure

Claude Sonnet 4.6 (Non-reasoning, High Effort) on Microsoft Azure

Performance benchmark now available
Claude Sonnet 4.6 (Non-reasoning, High Effort) on Google

Claude Sonnet 4.6 (Non-reasoning, High Effort) on Google

Performance benchmark now available
Claude Sonnet 4.6 (Non-reasoning, High Effort) on Anthropic

Claude Sonnet 4.6 (Non-reasoning, High Effort) on Anthropic

Performance benchmark now available
New Speech to Text model evaluation results available
Universal-3 Pro, AssemblyAI

Universal-3 Pro, AssemblyAI

17 Feb 2026

🔔 New article published
Claude Sonnet 4.6 - New leader in GDPval-AA

Claude Sonnet 4.6 - New leader in GDPval-AA

Claude Sonnet 4.6 wins in agentic performance of real-life knowledge work

🔔 New article published
Qwen3.5-397B-A17B - Everything you need to know

Qwen3.5-397B-A17B - Everything you need to know

Independent analysis and benchmarks of Qwen3.5-397B-A17B

New language model evaluation results available
Claude Sonnet 4.6 (Non-reasoning, High Effort)

Claude Sonnet 4.6 (Non-reasoning, High Effort)

Intelligence Index: 44 | AA-Omniscience: -3

New language model evaluation results available
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Intelligence Index: 52 | AA-Omniscience: 12

New language model evaluation results available
Qwen3.5 397B A17B (Reasoning)

Qwen3.5 397B A17B (Reasoning)

Intelligence Index: 45 | AA-Omniscience: -30

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Anthropic

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Anthropic

Performance benchmark now available
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Microsoft Azure

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Microsoft Azure

Performance benchmark now available
Qwen3.5 397B A17B (Reasoning) on Together.ai

Qwen3.5 397B A17B (Reasoning) on Together.ai

Performance benchmark now available
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Google

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Google

Performance benchmark now available
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Amazon Bedrock

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) on Amazon Bedrock

Performance benchmark now available
GLM-4.7-Flash (Non-reasoning) on Novita

GLM-4.7-Flash (Non-reasoning) on Novita

Performance benchmark now available
GLM-5 (Reasoning) on Parasail

GLM-5 (Reasoning) on Parasail

Performance benchmark now available
GLM-5 (Non-reasoning) on DeepInfra

GLM-5 (Non-reasoning) on DeepInfra

Performance benchmark now available
GLM-5 (Non-reasoning) on SiliconFlow

GLM-5 (Non-reasoning) on SiliconFlow

Performance benchmark now available
MiniMax-M2.5 on DeepInfra

MiniMax-M2.5 on DeepInfra

Performance benchmark now available
GLM-4.7 (Reasoning) on Clarifai

GLM-4.7 (Reasoning) on Clarifai

Performance benchmark now available
MiniMax-M2.5 on Clarifai

MiniMax-M2.5 on Clarifai

Performance benchmark now available
Qwen3.5 397B A17B (Reasoning) on Novita

Qwen3.5 397B A17B (Reasoning) on Novita

Performance benchmark now available
Qwen3.5 397B A17B (Non-reasoning) on Alibaba Cloud

Qwen3.5 397B A17B (Non-reasoning) on Alibaba Cloud

Performance benchmark now available
Qwen3.5 397B A17B (Reasoning) on Alibaba Cloud

Qwen3.5 397B A17B (Reasoning) on Alibaba Cloud

Performance benchmark now available

16 Feb 2026

13 Feb 2026

12 Feb 2026

11 Feb 2026

🔔 New article published
GLM-5 - Everything you need to know

GLM-5 - Everything you need to know

Independent analysis and benchmarks of GLM-5

New language model evaluation results available
GLM-5 (Reasoning)

GLM-5 (Reasoning)

Intelligence Index: 50 | AA-Omniscience: 2

GLM-5 (Reasoning) on DeepInfra

GLM-5 (Reasoning) on DeepInfra

Performance benchmark now available
GLM-5 (Reasoning) on Novita

GLM-5 (Reasoning) on Novita

Performance benchmark now available
New model in Text to Image Leaderboard
grok-imagine-image

grok-imagine-image

Text to Image ELO: 1172

New model in Text to Image Leaderboard
grok-imagine-image-pro

grok-imagine-image-pro

Text to Image ELO: 1137

New model in Image Editing Leaderboard
grok-imagine-image

grok-imagine-image

Image Editing ELO: 1224

New model in Image Editing Leaderboard
grok-imagine-image-pro

grok-imagine-image-pro

Image Editing ELO: 1213

New model in Text to Video Leaderboard
Veo 3.1 Fast

Veo 3.1 Fast

Text to Video ELO: 1209

New model in Text to Video Leaderboard
Veo 3.1

Veo 3.1

Text to Video ELO: 1209

New model in Text to Video Leaderboard
Kling 3.0 1080p (Pro)

Kling 3.0 1080p (Pro)

Text to Video ELO: 1242

New model in Text to Video Leaderboard
Kling 3.0 720p (Standard)

Kling 3.0 720p (Standard)

Text to Video ELO: 1214

New model in Image to Video Leaderboard
Veo 3.1 Fast

Veo 3.1 Fast

Image to Video ELO: 1288

New model in Image to Video Leaderboard
Veo 3.1

Veo 3.1

Image to Video ELO: 1249

New model in Image to Video Leaderboard
Kling 3.0 1080p (Pro)

Kling 3.0 1080p (Pro)

Image to Video ELO: 1277

New model in Image to Video Leaderboard
Kling 3.0 720p (Standard)

Kling 3.0 720p (Standard)

Image to Video ELO: 1268

10 Feb 2026

08 Feb 2026

06 Feb 2026

05 Feb 2026

🔔 New article published
Opus 4.6 Takes Lead in Agentic Real-World Knowledge Tasks

Opus 4.6 Takes Lead in Agentic Real-World Knowledge Tasks

We worked with Anthropic to benchmark Claude Opus 4.6 ahead of launch - it reached an Elo of 1606 with adaptive thinking, nearly 150 points ahead of GPT-5.2 (xhigh). This implies a win rate of ~70% when compared head-to-head with OpenAI's December 2025 flagship.

New language model evaluation results available
Claude Opus 4.6 (Adaptive Reasoning, Max Effort)

Claude Opus 4.6 (Adaptive Reasoning, Max Effort)

Intelligence Index: 53 | AA-Omniscience: 14

New language model evaluation results available
Claude Opus 4.6 (Non-reasoning, High Effort)

Claude Opus 4.6 (Non-reasoning, High Effort)

Intelligence Index: 46 | AA-Omniscience: 3

Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Amazon Bedrock

Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Amazon Bedrock

Performance benchmark now available
Claude Opus 4.6 (Non-reasoning, High Effort) on Google

Claude Opus 4.6 (Non-reasoning, High Effort) on Google

Performance benchmark now available
Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Google

Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Google

Performance benchmark now available
Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Microsoft Azure

Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Microsoft Azure

Performance benchmark now available
Claude Opus 4.6 (Non-reasoning, High Effort) on Microsoft Azure

Claude Opus 4.6 (Non-reasoning, High Effort) on Microsoft Azure

Performance benchmark now available
Qwen3 Coder Next on Parasail

Qwen3 Coder Next on Parasail

Performance benchmark now available
Claude Opus 4.6 (Non-reasoning, High Effort) on Anthropic

Claude Opus 4.6 (Non-reasoning, High Effort) on Anthropic

Performance benchmark now available
Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Anthropic

Claude Opus 4.6 (Adaptive Reasoning, Max Effort) on Anthropic

Performance benchmark now available
Claude Opus 4.6 (Non-reasoning, High Effort) on Amazon Bedrock

Claude Opus 4.6 (Non-reasoning, High Effort) on Amazon Bedrock

Performance benchmark now available
Qwen3 Coder Next on Novita

Qwen3 Coder Next on Novita

Performance benchmark now available
Qwen3 Coder Next on Together.ai

Qwen3 Coder Next on Together.ai

Performance benchmark now available
New model in Text to Image Leaderboard
Eigen Image

Eigen Image

Text to Image ELO: 1106

New model in Image Editing Leaderboard
Eigen Image

Eigen Image

Image Editing ELO: 1164

New model in Text to Video Leaderboard
PixVerse V5.6 (January)

PixVerse V5.6 (January)

Text to Video ELO: 1130

New model in Image to Video Leaderboard
PixVerse V5.6 (January)

PixVerse V5.6 (January)

Image to Video ELO: 1250

04 Feb 2026

03 Feb 2026

29 Jan 2026

28 Jan 2026

27 Jan 2026

🔔 New article published
Kimi K2.5 - Everything you need to know

Kimi K2.5 - Everything you need to know

Independent analysis and benchmarks of Moonshot's Kimi K2.5

New language model evaluation results available
Qwen3 Max Thinking

Qwen3 Max Thinking

Intelligence Index: 40 | AA-Omniscience: -34

New language model evaluation results available
Step3 VL 10B

Step3 VL 10B

Intelligence Index: 15 | AA-Omniscience: -59

New language model evaluation results available
Kimi K2.5 (Reasoning)

Kimi K2.5 (Reasoning)

Intelligence Index: 47 | AA-Omniscience: -8

New language model evaluation results available
K2 Think V2

K2 Think V2

Intelligence Index: 24 | AA-Omniscience: -34

Qwen3 Max Thinking on Alibaba Cloud

Qwen3 Max Thinking on Alibaba Cloud

Performance benchmark now available
Kimi K2.5 (Reasoning) on Kimi

Kimi K2.5 (Reasoning) on Kimi

Performance benchmark now available
Kimi K2.5 (Reasoning) on Novita

Kimi K2.5 (Reasoning) on Novita

Performance benchmark now available
Kimi K2.5 (Reasoning) on Fireworks

Kimi K2.5 (Reasoning) on Fireworks

Performance benchmark now available

26 Jan 2026

23 Jan 2026

New language model evaluation results available
LFM2 1.2B

LFM2 1.2B

Intelligence Index: 6 | AA-Omniscience: -72

GLM-4.7 (Reasoning) on Nebius

GLM-4.7 (Reasoning) on Nebius

Performance benchmark now available
DeepSeek V3.2 (Reasoning) on Nebius

DeepSeek V3.2 (Reasoning) on Nebius

Performance benchmark now available
New model in Text to Video Leaderboard
Kling 2.6 Standard (January)

Kling 2.6 Standard (January)

Text to Video ELO: 1190

New model in Text to Video Leaderboard
Kling 2.6 Pro (January)

Kling 2.6 Pro (January)

Text to Video ELO: 1196

New model in Text to Video Leaderboard
Kling O1 Pro (January)

Kling O1 Pro (January)

Text to Video ELO: 1203

New model in Text to Video Leaderboard
Kling O1 Standard (January)

Kling O1 Standard (January)

Text to Video ELO: 1185

New model in Image to Video Leaderboard
TeleVideo 2.0

TeleVideo 2.0

Image to Video ELO: 1262

New model in Image to Video Leaderboard
Kling 2.6 Pro (January)

Kling 2.6 Pro (January)

Image to Video ELO: 1254

New model in Image to Video Leaderboard
Kling 2.6 Standard (January)

Kling 2.6 Standard (January)

Image to Video ELO: 1281

22 Jan 2026

21 Jan 2026

20 Jan 2026

19 Jan 2026

16 Jan 2026

15 Jan 2026

New language model evaluation results available
Ring-1T

Ring-1T

Intelligence Index: 23 | AA-Omniscience: -48

New language model evaluation results available
Olmo 3.1 32B Instruct

Olmo 3.1 32B Instruct

Intelligence Index: 12 | AA-Omniscience: -51

MiniMax-M2.1 on Nebius

MiniMax-M2.1 on Nebius

Performance benchmark now available
Olmo 3.1 32B Instruct on DeepInfra

Olmo 3.1 32B Instruct on DeepInfra

Performance benchmark now available
New model in Text to Image Leaderboard
GLM-Image

GLM-Image

Text to Image ELO: 1048

New model in Text to Image Leaderboard
FLUX.2 [klein] 4B

FLUX.2 [klein] 4B

Text to Image ELO: 1069

New model in Text to Image Leaderboard
FLUX.2 [klein] 9B

FLUX.2 [klein] 9B

Text to Image ELO: 1134

New model in Text to Image Leaderboard
FLUX.2 [klein] Base 4B

FLUX.2 [klein] Base 4B

Text to Image ELO: 981

New model in Text to Image Leaderboard
FLUX.2 [klein] Base 9B

FLUX.2 [klein] Base 9B

Text to Image ELO: 1077

New model in Image Editing Leaderboard
GLM-Image

GLM-Image

Image Editing ELO: 929

New model in Image Editing Leaderboard
FLUX.2 [klein] 4B

FLUX.2 [klein] 4B

Image Editing ELO: 1108

New model in Image Editing Leaderboard
FLUX.2 [klein] 9B

FLUX.2 [klein] 9B

Image Editing ELO: 1166

New model in Image Editing Leaderboard
FLUX.2 [klein] Base 4B

FLUX.2 [klein] Base 4B

Image Editing ELO: 1029

New model in Image Editing Leaderboard
FLUX.2 [klein] Base 9B

FLUX.2 [klein] Base 9B

Image Editing ELO: 1127

Qwen Image on FlyMyAI

Qwen Image on FlyMyAI

Performance benchmark now available
FLUX.1 [dev] on FlyMyAI

FLUX.1 [dev] on FlyMyAI

Performance benchmark now available
FLUX.1 [schnell] on FlyMyAI

FLUX.1 [schnell] on FlyMyAI

Performance benchmark now available
New model in Text to Video Leaderboard
HunyuanVideo-1.5 (Fal)

HunyuanVideo-1.5 (Fal)

Text to Video ELO: 1015

New model in Image to Video Leaderboard
HunyuanVideo-1.5 (Fal)

HunyuanVideo-1.5 (Fal)

Image to Video ELO: 1118

14 Jan 2026

13 Jan 2026

12 Jan 2026

Loading more entries...