Changelog

Track all updates to LLMPodium: new models, benchmark updates, and feature releases.

2026-09-12

model

Added Agnes 3.0 Flash & DeepSeek V4.1 Flash

Added Sapiens AI Agnes 3.0 Flash (234 t/s, 1M context) and DeepSeek V4.1 Flash (552B MoE, 16B active, MIT open-weights reasoning).

2026-09-04

model

Added GPT-6 Astra

Added OpenAI GPT-6 Astra frontier reasoning model with 1.1M context, 100% ExploitBench, and Critical-tier autonomous cybersecurity safeguards.

2026-07-28

model

Added Qwen 3 235B

Added Alibaba Qwen 3 235B MoE model with hybrid thinking mode to all leaderboards.

2026-07-25

benchmark

LiveCodeBench scores updated

Updated LiveCodeBench scores for all models with latest contamination-free results.

2026-07-20

model

Added Llama 4 Maverick

Added Meta Llama 4 Maverick 400B MoE model with 1M context window support.

2026-07-15

update

Claude Opus 4 benchmark data

Added full benchmark suite for Claude Opus 4 including SWE-Bench and AIME scores.

2026-07-10

feature

Multi-language support

LLMPodium now available in 16 languages: EN, ZH, JA, KO, TH, RU, DE, ES, IT, FR, HI, AR, BN, PT, UR, ID.

2026-07-05

model

Added Gemini 2.5 Pro/Flash

Added Google Gemini 2.5 Pro and 2.5 Flash with thinking capabilities.

2026-07-01

update

Pricing data refresh

Updated pricing data for all models from official API documentation.

2026-06-28

benchmark

SWE-Bench Verified scores

Updated SWE-Bench Verified scores for frontier models.

2026-06-20

model

Added DeepSeek R1

Added DeepSeek R1 reasoning model with chain-of-thought capabilities.

2026-06-15

feature

Arena page launched

New Arena page for side-by-side model comparison launched.