BenchLeader

Claude Mythos 5.2 added, Claude Fable 5.2 up to #2 (10 Nov)

Claude Mythos 5.2 joins the board; Claude Fable 5.2 moved #3 → #2. 1 new model, 2 top-10 rank changes, 9 price changes, 82 new benchmark results.

BenchLeader refreshed 1695 models from 28 of 29 data feeds (benchmark publishers, aggregators and price lists). 1 new model, 2 top-10 rank changes, 9 price changes, 82 new or updated benchmark results across 28 data feeds.

New models

  • Claude Mythos 5.2 (Anthropic) is now listed at $16.00/M blended; no independent results yet. (details)

Movement in the top 10

  • Claude Fable 5.2 (high) moved from #3 to #2 in the BenchLeader Index. (details)
  • Claude Opus 5.5 (max) moved from #2 to #3 in the BenchLeader Index. (details)

Price changes

  • MiMo-V2.5 input price moved from $0.18/M to $0.11/M (39% down). (details)
  • MiMo-V2.5 output price moved from $0.36/M to $0.22/M (39% down). (details)
  • Nemotron 3.5 Lightning cost per task moved from $0.29/M to $0.31/M (9% up). (details)
  • MiMo-V2.5-Pro (thinking) input price moved from $0.44/M to $0.39/M (11% down). (details)
  • MiMo-V2.5-Pro (thinking) output price moved from $0.87/M to $0.78/M (10% down). (details)
  • MiMo-V2.5-Pro (no reasoning) input price moved from $0.67/M to $0.35/M (48% down). (details)
  • MiMo-V2.5-Pro (no reasoning) output price moved from $1.85/M to $0.70/M (62% down). (details)
  • MiMo-V2.5 (no reasoning) input price moved from $0.18/M to $0.11/M (39% down). (details)
  • MiMo-V2.5 (no reasoning) output price moved from $0.36/M to $0.22/M (39% down). (details)

New benchmark results

  • MLCR: 4 new results, including Claude Fable 5.2, Inkling 2 Small, Muse Ember… (board)
  • ForecastBench: 3 new results, including GPT-6 Astra, Muse Spark 1.3, Claude Fable 5.1 (board)
  • APEX-Agents: 3 new results, including Kimi K3.1 Pro, Inkling 2 Small, Gemini 4 Carbon (board)
  • SciCode: 3 new results, including Inkling 2 Small, Beam 1.0, DeepSeek V4.1 Flash 1017 (board)
  • ALE-Bench: 3 new results, including Grok 4.7, Qwen3.8 Max, Beam 1.0 (board)
  • Vals RSI Index: 2 new results, including Claude Fable 5.2, Gemini 4 Carbon (board)
  • Blueprint-Bench 2: 2 new results, including Muse Wonder, GPT-6.1 Sol (board)
  • Epoch Capabilities Index: 2 new results, including Kimi K3.1 Pro, Claude Fable 5.2 (board)
  • AA-Omniscience v1.1: 2 new results, including Apodex 1.1, gpt-oss-120b (board)
  • AA-Omniscience v1.1: accuracy: 2 new results, including Apodex 1.1, gpt-oss-120b (board)
  • AA-Omniscience v1.1: non-hallucination: 2 new results, including Apodex 1.1, gpt-oss-120b (board)
  • GPQA Diamond: 1 new result, including Claude Fable 5.2 (board)

Score revisions

  • Kimi K3 (max) on ForecastBench: 64.7% → 64.2%. (details)
  • GPT-5.6 Luna (xhigh) on OpenRouter token share: 1.4% → 0.8%. (details)
  • GPT-5.6 Luna (max) on OpenRouter token share: 1.4% → 0.8%. (details)
  • Hy4 on OpenRouter token share: 6.1% → 7.7%. (details)
  • GPT-5.6 Luna (high) on OpenRouter token share: 1.4% → 0.8%. (details)
  • Mistral Large 4 on Arena Hard Prompts: 1449 → 1455. (details)
  • Mistral Large 4 on Arena Coding: 1473 → 1482. (details)
  • Mistral Large 4 on Arena Maths: 1459 → 1465. (details)
  • Mistral Large 4 on Arena Instruction Following: 1414 → 1419. (details)
  • Mistral Large 4 on Arena Multi-turn: 1424 → 1430. (details)
  • Mistral Large 4 on Arena Longer Queries: 1418 → 1423. (details)
  • GPT-5.6 Luna (medium) on OpenRouter token share: 1.4% → 0.8%. (details)
  • GPT-5.6 Luna (low) on OpenRouter token share: 1.4% → 0.8%. (details)
  • GPT-5.6 Luna on OpenRouter token share: 1.4% → 0.8%. (details)
  • GPT-5.6 Luna (no reasoning) on OpenRouter token share: 1.4% → 0.8%. (details)
  • Dolly 2.0 12B on Epoch Capabilities Index: 56.4 → 57.1. (details)
  • Grok 4.7 on Design Arena Agentic: 1239 → 1233. (details)
  • Cerebras-GPT-13B on Epoch Capabilities Index: 49.9 → 50.5. (board)
  • CodeQwen1.5-7B on Epoch Capabilities Index: 75.6 → 76.3. (board)
  • DeepSeek Coder 6.7B on Epoch Capabilities Index: 68.3 → 68.8. (board)
  • …and 7 more.

Speed changes

  • Claude Fable 5.2 output speed changed from 72 tok/s to 103 tok/s. (details)
  • GPT-6 Astra time to first token changed from 3.52 s to 4.92 s. (details)
  • Claude Fable 5 time to first token changed from 5.41 s to 3.38 s. (details)
  • GPT-5.6 Sol time to first token changed from 2.24 s to 3.29 s. (details)
  • GPT-5.5 time to first answer changed from 44.08 s to 28.62 s. (details)
  • GPT-5.5 time to first token changed from 3.08 s to 1.92 s. (details)
  • Claude Opus 4.7 time to first token changed from 3.49 s to 1.02 s. (details)
  • GPT-6 Sol time to first answer changed from 72.11 s to 45.48 s. (details)
  • Kimi K2.6 time to first token changed from 1.12 s to 0.73 s. (details)
  • Grok 4.20 Multi-Agent time to first token changed from 16.72 s to 7.06 s. (details)
  • Grok 4 time to first answer changed from 13.40 s to 18.80 s. (details)
  • Grok 4.7 time to first token changed from 3.06 s to 1.94 s. (details)
  • GPT-5.2-Codex time to first token changed from 3.73 s to 1.86 s. (details)
  • DeepSeek V4 Flash time to first answer changed from 41.94 s to 27.04 s. (details)
  • DeepSeek V4 Flash thinking time changed from 40.14 s to 25.49 s. (details)
  • …and 86 more.

Source status

  • Source t3code failed to update: unexpected payload (schema 4).

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