Blog

Muse Wonder Full + 15 new AI models: 12 Oct–18 Oct

16 new models incl. Muse Wonder Full, Muse Wonder and DeepSeek V4.1 Flash 1017, Mimo v2 Omni up 1.3, 29 price cuts, 1400 new benchmark results: what moved on the BenchLeader Index this week.

Claude Haiku 5.5 + 3 new AI models: 5 Oct–11 Oct

4 new models incl. Claude Haiku 5.5, GPT-6 Sol (Daybreak Blue) and Ling 3.0 Flash Sante, Claude Sonnet 5.5 up 5.8, 13 price cuts, 620 new benchmark results: what moved on the BenchLeader Index this week.

Gemini 4 Argon, GPT-6.1 Sol + 8 new AI models: 28 Sep–4 Oct

10 new models incl. Gemini 4 Argon, GPT-6.1 Sol and Claude Sonnet 5.5, 5 price cuts, 613 new benchmark results: what moved on the BenchLeader Index this week.

Claude Opus 5.5, GPT-6 Sol + 10 new AI models: 21 Sep–27 Sep

12 new models incl. Claude Opus 5.5, GPT-6 Sol and MiMo-V2.6-Pro, GPT-6 Luna up 2.1, 23 price cuts, 2211 new benchmark results: what moved on the BenchLeader Index this week.

Step 5 Preview + 12 new AI models: 14 Sep–20 Sep

13 new models incl. Step 5 Preview, Mercury 2.5 and GLM 5.3 FlashX, DeepSeek V4 Flash up 10.0, 15 price cuts, 6954 new benchmark results: what moved on the BenchLeader Index this week.

agnes-3-0-flash: new AI model, 7 Sep–13 Sep

1 new model incl. agnes-3-0-flash, MiniMax M3 up 7.8, 37 price cuts, 3 new benchmark results: what moved on the BenchLeader Index this week.

The best AI models for coding, by the benchmarks that actually test coding

Which LLMs write and fix code best, judged by SWE-bench Verified, LiveCodeBench, Aider, Terminal-Bench, SWE-Bench Pro and WebDev Arena, with prices.

The fastest LLMs: output speed and time to first answer, measured

Which AI models stream fastest and answer soonest, measured on live traffic, and why reasoning effort turns a two-second model into a two-minute one.

LLM API pricing compared: what a task really costs in 2026

Per-token prices hide the shape of real work. What a chat reply, a document summary, a code edit and an agent session cost, with and without caching.

How to read an LLM leaderboard: ties, price, speed and reasoning effort

How to read AI model rankings: when a lead is real, why price and speed change the answer, and what reasoning effort does to a score.

Open weights vs closed models in 2026: how far behind, and where it doesn't matter

The best open-weights LLMs ranked against the closed frontier, with the gap in points, price and speed, from independent benchmarks.

Reasoning effort explained: why the same model gets two different scores

Low, medium, high, xhigh, max: what the reasoning-effort setting does to quality, speed and cost, and why a benchmark score without it is half a number.

Introducing BenchLeader: every AI benchmark, one leaderboard

Why we built an aggregator instead of another benchmark, how the BenchLeader Index works, and what you can do with the data.

Qwen3 30B A3B 2507 got cheaper; Gemini 3.7 Flash got cheaper. 99 price changes, 58 new benchmark results.

Muse Wonder Full moved #15 → #3; GLM 4.5V got cheaper. 8 top-10 rank changes, 85 price changes, 194 new benchmark results.

MiniMax-M3.1-Flash joins the board; Mercury 3 joins the board. 2 new models, 10 top-10 rank changes, 77 price changes, 384 new benchmark results.

Grok 4.20 got cheaper; DeepSeek V4 Pro got cheaper. 16 price changes, 812 new benchmark results.

Muse Spark Open 816B A42B joins the board; Beam 1.0 joins the board. 7 new models, 2 top-10 rank changes, 37 price changes, 78 new benchmark results.

Muse Spark Open joins the board; Nemotron 3.5 Bolt 74B-A9B joins the board. 4 new models, 290 new benchmark results.

Nemotron 3.5 Super VL 120B A12B joins the board. 1 new model, 8 price changes, 313 new benchmark results.

Musespark 1.3 joins the board; LFM2.5-24B-A2B joins the board. 2 new models, 4 top-10 rank changes, 3 price changes, 122 new benchmark results.

Muse Wonder joins the board; Muse Wonder Full joins the board. 2 new models, 3 top-10 rank changes, 86 price changes, 212 new benchmark results.

33 new or updated benchmark results across 23 data feeds.

91 new or updated benchmark results across 23 data feeds.

Nemotron 3 Nano 30B A3B got cheaper; deepseek-v4-flash-0731 got cheaper. 14 price changes, 50 new benchmark results.

GPT-6 Sol (Daybreak Blue) joins the board. 2 new models, 3 top-10 rank changes, 3 price changes, 151 new benchmark results.

Claude Opus 5.5 climbed from #10 to #1 on the BenchLeader Index. 5 new models, 11 top-10 rank changes, 26 price changes, 157 new benchmark results.

GPT-6 Astra moved #4 → #3; Nemotron 3 Nano Omni 30B A3B got cheaper. 2 top-10 rank changes, 3 price changes, 20 new benchmark results.

Claude Fable 5.1 climbed from #36 to #1 on the BenchLeader Index. 1 new model, 10 top-10 rank changes, 38 price changes, 299 new benchmark results.

Claude Opus 5.5 moved #7 → #2; Gemini 4 Argon moved #11 → #7. 8 top-10 rank changes, 6 price changes, 65 new benchmark results.

61 new or updated benchmark results across 23 data feeds.

Ling 3.1 Flash joins the board; Claude Opus 5.5 moved #4 → #3. 1 new model, 9 top-10 rank changes, 22 price changes, 149 new benchmark results.

Gemini 4 Argon joins the board; GPT-6.1 Sol joins the board. 3 new models, 5 top-10 rank changes, 19 price changes, 399 new benchmark results.

Claude Sonnet 5.5 joins the board; Claude Opus 5.5 Range joins the board. 4 new models, 9 top-10 rank changes, 20 price changes, 169 new benchmark results.

Daybreak Blue joins the board; Daybreak Red joins the board. 2 new models, 2 top-10 rank changes, 12 price changes, 93 new benchmark results.

15 price changes, 155 new or updated benchmark results across 22 data feeds.

114 new or updated benchmark results across 22 data feeds.

Claude Fable 5.1 moved #5 → #4. 2 top-10 rank changes, 70 new benchmark results.

GLM 5.3 Prime joins the board; Qwen3.8 Max Prime joins the board. 2 new models, 12 price changes, 21 new benchmark results.

GPT-6 Sol joins the board; GPT-6 Luna joins the board. 14 new models, 5 top-10 rank changes, 129 price changes, 83 new benchmark results.

GPT-6 Astra climbed from #2 to #1 on the BenchLeader Index. 8 new models, 9 top-10 rank changes, 48 price changes, 1940 new benchmark results.

Grok 4.7 joins the board; Dots3 Note joins the board. 3 new models, 2 top-10 rank changes, 4 price changes, 88 new benchmark results.

Muse Glimmer got cheaper. 9 price changes, 220 new benchmark results.

Claude Fable 5.1 climbed from #8 to #1 on the BenchLeader Index. 2 new models, 12 top-10 rank changes, 41 price changes, 6874 new benchmark results.

GPT-6 Astra moved #6 → #4; Claude Fable 5.1 moved #7 → #5. 8 top-10 rank changes, 13 new benchmark results.

union-alpha joins the board; Jamba Large joins the board. 3 new models, 4 top-10 rank changes, 17 price changes, 11 new benchmark results.

qwen3-8 joins the board; solar-pro-4 joins the board. 3 new models, 2 top-10 rank changes, 12 price changes, 23 new benchmark results.

Mercury 2.5 joins the board; Claude Opus 5 moved #9 → #8. 1 new model, 2 top-10 rank changes, 5 price changes, 13 new benchmark results.

k2-1b-final joins the board; k2-4b-ph1 joins the board. 4 new models, 6 top-10 rank changes, 3 price changes, 81 new benchmark results.

au-anthropic-claude-opus-4-6-v1 joins the board. 93 new models, 4 top-10 rank changes, 8 price changes, 33 new benchmark results.

Claude Fable 5.1 climbed from #7 to #1 on the BenchLeader Index. 10 new models, 9 top-10 rank changes, 6 price changes, 389 new benchmark results.

agnes-3-0-flash joins the board; LongCat 2.0 got cheaper. 1 new model, 4 price changes, 5 new benchmark results.

mimo-v2-omni-0327 joins the board; Qwen3 joins the board. 78 new models, 2 top-10 rank changes, 112 price changes, 223 new benchmark results.

GPT-6 Astra climbed from #2 to #1 on the BenchLeader Index. 261 new models, 10 top-10 rank changes, 122 price changes, 4579 new benchmark results.