BenchLeader

crouton-alpha added, Gemini Omni 1.1 Flash added (5 Nov)

crouton-alpha and Gemini Omni 1.1 Flash join the board. 4 new models, 1 price change, 302 new benchmark results. Daily digest for 5 Nov 2026.

BenchLeader refreshed 1692 models from 27 of 27 data feeds (benchmark publishers, aggregators and price lists). 4 new models, 1 price change, 302 new or updated benchmark results across 27 data feeds.

New models

  • Crouton Alpha (Unknown) is now listed; no independent results yet. (details)
  • Gemini Omni 1.1 Flash (Google) is now listed at $3.38/M blended; no independent results yet. (details)
  • Gemini Omni 1.2 Flash (Google) is now listed at $3.38/M blended; no independent results yet. (details)
  • Ripple-65B Preview (Ricursive) is now listed at $0.30/M blended; no independent results yet. (details)

Price changes

  • Gemini 4 Argon (high) cost per task moved from $7.68/M to $9.95/M (29% up). (details)

New benchmark results

  • AA Intelligence Index v5.0.1: 10 new results, including Claude Opus 5, Grok 4.6, Muse Wonder… (board)
  • CritPt v2 hard subset (AA): 10 new results, including Claude Opus 5, Muse Wonder, GPT-5.6 Luna… (board)
  • FrontierCode 2.0 (AA): 9 new results, including Claude Opus 5, Muse Wonder, Hy4… (board)
  • HLE-Diamond (AA): 6 new results, including Claude Opus 5, Grok 4.6, Muse Wonder… (board)
  • Terminal-Bench-Science 0.1 (AA): 6 new results, including Grok 4.6, Muse Wonder, Hy4… (board)
  • Chess Puzzles: 6 new results, including Doubao Seed 2.0 Lite 260428, Doubao Seed 2.0 Pro 260215, Doubao Seed 2.1 Pro 260628… (board)
  • FORTRESS: 5 new results, including GPT-6 Astra, GLM 5.3 Flash, Beam 1.0 501B A23B… (board)
  • FACTS Search: 5 new results, including Kimi K3, Qwen3.8 Max, gpt-oss-120b… (board)
  • AA-Omniscience v1.1: 4 new results, including Claude Opus 5, Muse Wonder, Hy4… (board)
  • AA-Omniscience v1.1: accuracy: 4 new results, including Claude Opus 5, Muse Wonder, Hy4… (board)
  • AA-Omniscience v1.1: non-hallucination: 4 new results, including Claude Opus 5, Muse Wonder, Hy4… (board)
  • AutomationBench: 3 new results, including Muse Wonder, Hy4, Gemma 4 12B (board)

Score revisions

  • Claude Opus 5.5 (max) on Coding-agent share: 54.4% → 55.4%. (details)
  • GPT-6 Astra on FACTS Search: 89.9% → 0.9%. (details)
  • GPT-6 Astra (max) on Arena Agent: 13.7 → 14.7. (details)
  • Claude Fable 5.1 (max) on Arena Agent: 13.1 → 13.9. (details)
  • Claude Fable 5 on Epoch Capabilities Index: 163 → 164.2. (details)
  • Claude Fable 5.1 on Epoch Capabilities Index: 166.4 → 166.9. (details)
  • Claude Opus 5.5 (xhigh) on Coding-agent share: 54.4% → 55.4%. (details)
  • Claude Opus 5.5 (high) on Coding-agent share: 54.4% → 55.4%. (details)
  • Claude Sonnet 5.5 (max) on Arena Agent: 14.6 → 15.4. (details)
  • GPT-6.1 Terra (max) on Arena Agent: 10.2 → 11.4. (details)
  • Muse Spark on Epoch Capabilities Index: 152.2 → 153.1. (details)
  • GPT-6 Sol (max) on CritPt v2 hard subset (AA): 66% → 66.5%. (details)
  • Claude Opus 5.5 on FACTS Search: 86.5% → 0.9%. (details)
  • Claude Opus 5.5 on Coding-agent share: 54.4% → 55.4%. (details)
  • Claude Opus 4.7 on Epoch Capabilities Index: 156.7 → 157.6. (details)
  • GPT-6.1 Sol on FACTS Search: 88.2% → 0.9%. (details)
  • GPT-6.1 Sol on Epoch Capabilities Index: 167.8 → 168.9. (details)
  • Muse Spark 1.1 on Epoch Capabilities Index: 154.2 → 154.9. (details)
  • Claude Opus 5.5 (medium) on Coding-agent share: 54.4% → 55.4%. (details)
  • GPT-5.5 on Epoch Capabilities Index: 159.4 → 160. (details)
  • …and 200 more.

Speed changes

  • GPT-6 Astra output speed changed from 25 tok/s to 40 tok/s. (details)
  • GPT-5.5 time to first token changed from 1.56 s to 2.33 s. (details)
  • GPT-5.4 Pro output speed changed from 5 tok/s to 7 tok/s. (details)
  • Kimi K3 time to first token changed from 2.92 s to 1.14 s. (details)
  • Kimi K2.6 time to first token changed from 0.50 s to 0.69 s. (details)
  • Kimi K3 output speed changed from 30 tok/s to 63 tok/s. (details)
  • Grok 4.20 Multi-Agent time to first token changed from 3.56 s to 4.96 s. (details)
  • DeepSeek V4 Flash time to first token changed from 1.86 s to 0.99 s. (details)
  • GPT-5.2-Codex time to first token changed from 4.19 s to 2.25 s. (details)
  • Qwen3.5 397B A17B time to first token changed from 0.50 s to 1.03 s. (details)
  • Qwen3.6 27B time to first token changed from 2.03 s to 0.64 s. (details)
  • GLM 5.1 time to first token changed from 2.19 s to 1.29 s. (details)
  • GLM 5V Turbo output speed changed from 31 tok/s to 46 tok/s. (details)
  • GLM 5V Turbo time to first token changed from 1.05 s to 1.47 s. (details)
  • Hy4 output speed changed from 33 tok/s to 46 tok/s. (details)
  • …and 55 more.

Today's top five

  1. Claude Fable 5.2 (max) — 72.7
  2. Gemini 4 Carbon (high) — 68.9
  3. Claude Opus 5.5 (max) — 68.4
  4. GPT-6 Astra (max) — 67.3
  5. Claude Fable 5.1 (max) — 66.7

More: New AI model releases this week, and the full LLM leaderboard 2026.

This digest is generated automatically from the day's data changes. Every line links to the page where you can check the numbers and their source.

Head to head

Get this in your inbox

No ads, no tracking pixels, unsubscribe in one click.
What to receive