Recap · 2026-10

AI leaderboard changes October 2026, so far

As of Oct 14, 2026, 27 confirmed crown changes are recorded for October 2026.

Crowns

Confirmed crown changes

Oct 14, 2026. Decision models: vLLM-SR d3 (27B) and linne-oak 1.0 (31B) and Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) are co-SOTA. Primary sources disagree beyond the margin of error, so the crown is shared with low confidence. vLLM-SR d3 (27B)linne-oak 1.0 (31B)Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) · Decision models
Oct 14, 2026. Decision models: vLLM-SR d3 (27B) and linne-oak 1.0 (31B) and Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) are co-SOTA. Primary sources disagree beyond the margin of error, so the crown is shared with low confidence. vLLM-SR d3 (27B)linne-oak 1.0 (31B)Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) · Decision models
Oct 14, 2026. Decision models: vLLM-SR d3 (27B) and linne-oak 1.0 (31B) and Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) are co-SOTA. Primary sources disagree beyond the margin of error, so the crown is shared with low confidence. vLLM-SR d3 (27B)linne-oak 1.0 (31B)Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) · Decision models
Oct 14, 2026. Decision models: Bobcat 1.3 (31B) takes the crown. First holder recorded for this category and slice. Bobcat 1.3 (31B) · Decision models
Oct 14, 2026. Decision models › Vision: RSI-Jev v6.1 VL 27B (effort auto) takes the crown. First holder recorded for this category and slice. RSI-Jev v6.1 VL 27B (effort auto) · Decision models › Vision
Oct 14, 2026. Decision models: Nace.AI Drex v2.0 (35B-A3B) takes the crown. First holder recorded for this category and slice. Nace.AI Drex v2.0 (35B-A3B) · Decision models
Oct 13, 2026. Reasoning: Kimi K3 and Mimo V2.6 Pro are co-SOTA. Crown change confirmed by two consecutive snapshots or two independent primary sources. Kimi K3 Mimo V2.6 Pro · Reasoning
Oct 13, 2026. Long context: Kimi K3 and Mimo V2.6 Pro are co-SOTA. Crown change confirmed by two consecutive snapshots or two independent primary sources. Kimi K3 Mimo V2.6 Pro · Long context
Oct 13, 2026. General LLM: Kimi K3 and Mimo V2.6 Pro and GLM 5.3 are co-SOTA. Crown change confirmed by two consecutive snapshots or two independent primary sources. Kimi K3 Mimo V2.6 Pro GLM 5.3 · General LLM
Oct 8, 2026. Decision models: decisio v0.8.0 on gemma-4-31B-it (frozen, FP8 on load, one prefill per state, self-hosted) takes the crown. First holder recorded for this category and slice. decisio v0.8.0 on gemma-4-31B-it (frozen, FP8 on load, one prefill per state, self-hosted) · Decision models
Oct 8, 2026. Decision models: Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) takes the crown. First holder recorded for this category and slice. Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) · Decision models
Oct 8, 2026. Decision models: Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) takes the crown. First holder recorded for this category and slice. Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) · Decision models
Oct 8, 2026. Reasoning: Gemma 4 31B and Qwen3 8 27B and Gemma 4 26B A4b are co-SOTA. First holder recorded for this category and slice. Gemma 4 31B Qwen3 8 27B Gemma 4 26B A4b · Reasoning
Oct 8, 2026. Long context: Gemma 4 31B and Gemma 4 26B A4b are co-SOTA. First holder recorded for this category and slice. Gemma 4 31B Gemma 4 26B A4b · Long context
Oct 8, 2026. General LLM: Gemma 4 31B and Gemma 4 26B A4b are co-SOTA. First holder recorded for this category and slice. Gemma 4 31B Gemma 4 26B A4b · General LLM
Oct 8, 2026. Factuality: Gemma 4 31B and Gemma 4 26B A4b and Qwen3 8 27B are co-SOTA. First holder recorded for this category and slice. Gemma 4 31B Gemma 4 26B A4b Qwen3 8 27B · Factuality
Oct 8, 2026. Decision models: Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) takes the crown. First holder recorded for this category and slice. Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) · Decision models
Oct 8, 2026. Decision models: Sage 1.3.0 (Levanto Labs) takes the crown. First holder recorded for this category and slice. Sage 1.3.0 (Levanto Labs) · Decision models
Oct 6, 2026. Coding › Agentic: Claude Opus 5.5 and Claude Sonnet 5.5 are co-SOTA. First holder recorded for this category and slice. Claude Opus 5.5 Claude Sonnet 5.5 · Coding › Agentic
Oct 6, 2026. Coding: Claude Opus 5.5 and Claude Sonnet 5.5 are co-SOTA. First holder recorded for this category and slice. Claude Opus 5.5 Claude Sonnet 5.5 · Coding
Oct 6, 2026. Coding › Agentic: GLM 5.3 and GLM 5.3 Flash are co-SOTA. First holder recorded for this category and slice. GLM 5.3 GLM 5.3 Flash · Coding › Agentic
Oct 6, 2026. Coding: GLM 5.3 and GLM 5.3 Flash are co-SOTA. First holder recorded for this category and slice. GLM 5.3 GLM 5.3 Flash · Coding
Oct 6, 2026. Image edit: Flux 2 Klein 4B takes the crown. First holder recorded for this category and slice. Flux 2 Klein 4B · Image edit
Oct 2, 2026. Reasoning: Gemini 4 Argon and Claude Opus 5.5 are co-SOTA. Crown change confirmed by two consecutive snapshots or two independent primary sources. Gemini 4 Argon Claude Opus 5.5 · Reasoning
Oct 2, 2026. Long context: Gemini 4 Argon and Claude Opus 5.5 are co-SOTA. Crown change confirmed by two consecutive snapshots or two independent primary sources. Gemini 4 Argon Claude Opus 5.5 · Long context
Oct 2, 2026. General LLM: Gemini 4 Argon takes the crown. Crown change confirmed by two consecutive snapshots or two independent primary sources. Gemini 4 Argon · General LLM
Oct 2, 2026. Factuality: Gemini 4 Argon and Claude Opus 5.5 and Claude Fable 5.1 are co-SOTA. Crown change confirmed by two consecutive snapshots or two independent primary sources. Gemini 4 Argon Claude Opus 5.5 Claude Fable 5.1 · Factuality

Top 3

Models that entered the top 3

Claude Fable 5 at #1 in Abstraction (ARC)
Claude Fable 5.1 at #1 in Factuality
Claude Opus 5.5 at #1 in Reasoning
Claude Opus 5.5 at #1 in Coding
Claude Opus 5.5 at #1 in Coding › Agentic
Claude Opus 5.5 at #1 in Long context
Claude Opus 5.5 at #1 in Factuality
Claude Opus 5.5 at #1 in Abstraction (ARC)
Claude Sonnet 5.5 at #1 in Coding
Claude Sonnet 5.5 at #1 in Coding › Agentic
Gemini 3.8 Flash at #1 in Abstraction (ARC)
Gemini 4 Argon at #1 in General LLM
Gemini 4 Argon at #1 in Reasoning
Gemini 4 Argon at #1 in Science
Gemini 4 Argon at #1 in Long context
Gemini 4 Argon at #1 in Factuality
GPT-6 Astra at #1 in Abstraction (ARC)
GPT-6.1 Sol at #1 in Abstraction (ARC)
Jcorners Ingot 8B R3 at #1 in Embeddings
linne-oak 1.0 (31B) at #1 in Decision models
Muse Wonder at #1 in General LLM
Muse Wonder at #1 in Reasoning
Muse Wonder at #1 in Factuality
Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) at #1 in Decision models
vLLM-SR d3 (27B) at #1 in Decision models
Bytedance Seed1 6 Embedding 1215 at #2 in Embeddings
Claude Fable 5.1 at #2 in Math
Claude Opus 4.6 at #2 in Reasoning
Claude Opus 4.6 at #2 in Long context
Claude Opus 5.5 at #2 in General LLM
Decider 31B (NVFP4) at #2 in Decision models
Gemini 3.1 Pro Preview at #2 in Abstraction (ARC)
GPT Image 2.5 Flare at #2 in Image generation
GPT Image 2.5 Flare at #2 in Image edit
GPT-6 Astra at #2 in Science
GPT-6 Astra at #2 in Coding
GPT-6 Astra at #2 in Coding › Agentic
GPT-6 Astra at #2 in Coding › Web dev
GPT-6.1 Sol at #2 in Coding
GPT-6.1 Sol at #2 in Coding › Agentic
Kingsoft LLM Qzhou Embedding at #2 in Embeddings
Muse Wonder at #2 in Vision
Veo 3 at #2 in Video generation
Claude Fable 5 at #3 in General LLM
Claude Fable 5 at #3 in Reasoning
Claude Fable 5 at #3 in Long context
Claude Fable 5.1 at #3 in Coding
Claude Fable 5.1 at #3 in Coding › Agentic
Claude Fable 5.1 at #3 in Abstraction (ARC)
Claude Opus 4.6 at #3 in Factuality
Claude Opus 5 at #3 in Abstraction (ARC)
Claude Sonnet 5.5 at #3 in Science
Claude Sonnet 5.5 at #3 in Coding › Web dev
Gemini 3.1 Pro Preview at #3 in Math
GPT Image 2 at #3 in Image generation
GPT Image 2 at #3 in Image edit
GPT-5.6 Sol at #3 in Vision
GPT-5.6 Sol at #3 in Abstraction (ARC)
Kandinsky 5.0 T2v Pro at #3 in Video generation
Llama 3.1 405B Instruct at #3 in General LLM
Nvidia Llama Embed Nemotron 8B at #3 in Embeddings
Sieve 31B at #3 in Decision models

Benchmarks

Benchmarks added or saturated

Benchmark additions for this month are still empty.