Show HN: Independent LLM benchmark index tracking 22 frontier models

Empirical evaluations measuring coding proficiency, reasoning benchmarks (GPQA, SWE-bench Verified, HLE), and real-time value-for-money ($/point) metrics.

Artificial Analysis Intelligence Index • Higher is better

Output tokens per second • Higher is better

Weighted average blended price (USD) • Lower is better

Intelligence Index v4.3 incorporates evaluations: SWE-bench Verified (40%), Terminal-Bench 2.0 (25%), Aider Polyglot (20%), LiveCodeBench (15%). Every figure is verified against primary source logs without prompt leakage.

Code Intelligence Index vs. Weighted average cost (USD, Log Scale)

Muse Glimmer (high) (Meta) Intelligence Index: 18.6 Blended Price: $0.12/1M Speed: 93 tps

Muse Glimmer..

Gemini 3.5 Flash-Lite (Google) Intelligence Index: 23.2 Blended Price: $0.17/1M Speed: 358 tps

Gemini 3.5 F..

gpt-oss-120b (high) (OpenAI) Intelligence Index: 12.6 Blended Price: $0.18/1M Speed: 181 tps

gpt-oss-120b..

GLM-5.3-Flash (Z AI) Intelligence Index: 42.1 Blended Price: $0.35/1M Speed: 95 tps

DeepSeek V4.1 Flash (max) (DeepSeek) Intelligence Index: 40.0 Blended Price: $0.37/1M Speed: 208 tps

Qwen3.8-Flash-Next (Alibaba) Intelligence Index: 40.0 Blended Price: $0.45/1M Speed: 55 tps

DeepSeek V4 Flash Vision (max) (DeepSeek) Intelligence Index: 35.0 Blended Price: $0.47/1M Speed: 214 tps

GPT-5.6 Luna (max) (OpenAI) Intelligence Index: 38.0 Blended Price: $0.52/1M Speed: 133 tps

Mistral Medium 3.5 (Mistral) Intelligence Index: 15.6 Blended Price: $0.60/1M Speed: 137 tps

MiniMax-M3 (MiniMax) Intelligence Index: 30.6 Blended Price: $0.60/1M Speed: 186 tps

Nemotron 3 Ultra (NVIDIA) Intelligence Index: 23.0 Blended Price: $0.75/1M Speed: 164 tps

Inkling (Thinking Machines) Intelligence Index: 26.6 Blended Price: $0.75/1M Speed: 81 tps

Qwen3.8 27B (xhigh) (Alibaba) Intelligence Index: 34.0 Blended Price: $0.90/1M Speed: 43 tps

DeepSeek V4 Pro Benchmarks, Price & Specs (Oct 2026) (DeepSeek) Intelligence Index: 36.8 Blended Price: $0.96/1M Speed: 88 tps

Step 5 Preview (StepFun) Intelligence Index: 44.1 Blended Price: $1.20/1M Speed: 100 tps

K2 Horizon 375B A23B (Institute of Foundation Models) Intelligence Index: 31.8 Blended Price: $1.20/1M Speed: 80 tps

Gemini 3.8 Flash (high) (Google) Intelligence Index: 41.0 Blended Price: $1.31/1M Speed: 298 tps

Gemini 3.1 Pro Benchmarks, Price & Specs (Oct 2026) (Google) Intelligence Index: 30.2 Blended Price: $1.40/1M Speed: 119 tps

GLM-5.3 (max) (Z AI) Intelligence Index: 45.0 Blended Price: $1.50/1M Speed: 72 tps

Kimi K3 (max) (Kimi) Intelligence Index: 44.0 Blended Price: $1.80/1M Speed: 38 tps

Muse Spark 1.3 (max) (Meta) Intelligence Index: 48.0 Blended Price: $2.25/1M Speed: 214 tps

Muse Spark 1..

Qwen3.8 2.4T A95B (Alibaba) Intelligence Index: 40.0 Blended Price: $2.25/1M Speed: 38 tps

Qwen3.8 Max (0902) (Alibaba) Intelligence Index: 45.0 Blended Price: $3.00/1M Speed: 37 tps

GPT-5.6 Terra (max) (OpenAI) Intelligence Index: 42.1 Blended Price: $3.15/1M Speed: 84 tps

Grok 4.6 (high) (SpaceXAI) Intelligence Index: 44.0 Blended Price: $4.00/1M Speed: 60 tps

GPT-5.6 Sol (max) Benchmarks, Price & Specs (Oct 2026) (OpenAI) Intelligence Index: 48.0 Blended Price: $6.12/1M Speed: 61 tps

GPT-5.6 Sol ..

Claude Sonnet 5 (max) (Anthropic) Intelligence Index: 38.1 Blended Price: $8.00/1M Speed: 74 tps

GPT-6 Astra (max) Benchmarks, Price & Specs (Oct 2026) (OpenAI) Intelligence Index: 55.0 Blended Price: $12.00/1M Speed: 57 tps

GPT-6 Astra ..

Claude Opus 5 (max) Benchmarks, Price & Specs (Oct 2026) (Anthropic) Intelligence Index: 52.6 Blended Price: $20.00/1M Speed: 51 tps

Claude Opus ..

Claude Fable 5.1 (max with fallback) (Anthropic) Intelligence Index: 53.1 Blended Price: $30.00/1M Speed: 68 tps

Claude Fable..

Frontier Language Model Intelligence, Over Time

Progressive evolution across Anthropic, OpenAI, DeepSeek, Google, and Meta from 2023 to 2026.

Claude 3 Opus (Anthropic) Date: 2024-03-04 Score: 35

Claude 3.5 Sonnet (Anthropic) Date: 2024-06-20 Score: 45

Claude 3.5 Sonnet (New) (Anthropic) Date: 2024-10-22 Score: 54

Claude 3.7 Sonnet (Anthropic) Date: 2025-02-24 Score: 70

GPT-4 Turbo (OpenAI) Date: 2023-11-06 Score: 28

GPT-4o (OpenAI) Date: 2024-05-13 Score: 38

o1-preview (OpenAI) Date: 2024-09-12 Score: 45

o1 (OpenAI) Date: 2024-12-17 Score: 49

o3-mini (OpenAI) Date: 2025-01-31 Score: 51

DeepSeek Coder 6.7B (DeepSeek) Date: 2023-11-15 Score: 15

DeepSeek V2 (DeepSeek) Date: 2024-05-06 Score: 28

DeepSeek V3 (DeepSeek) Date: 2024-12-26 Score: 42

DeepSeek R1 (DeepSeek) Date: 2025-01-20 Score: 49

Gemini 1.5 Pro (Google) Date: 2024-02-15 Score: 32

Gemini 2.0 Flash Thinking (Google) Date: 2024-12-19 Score: 57

Gemini 2.0 Pro (Google) Date: 2025-02-05 Score: 52

Llama 3 70B (Meta) Date: 2024-04-18 Score: 24

Llama 3.1 405B (Meta) Date: 2024-07-23 Score: 38

Llama 3.3 70B (Meta) Date: 2024-12-06 Score: 42

Canonical Benchmark Suite

Breakdown across isolated execution harnesses matching Artificial Analysis methodologies.