Daily digest, 5 Oct 2026

BenchLeader refreshed 1550 models from 23 of 23 sources. 8 top-10 rank changes, 6 price changes, 65 new or updated benchmark results across 23 sources.

Movement in the top 10

• Claude Opus 5.5 (high) moved from #7 to #2 in the BenchLeader Index. (details)
• Claude Opus 5.5 (thinking) moved from #2 to #3 in the BenchLeader Index. (details)
• Claude Opus 5.5 (max) moved from #3 to #4 in the BenchLeader Index. (details)
• Claude Fable 5.1 (xhigh) moved from #4 to #6 in the BenchLeader Index. (details)
• Gemini 4 Argon (high) moved from #11 to #7 in the BenchLeader Index. (details)
• Claude Fable 5.1 (high) moved from #6 to #8 in the BenchLeader Index. (details)
• Claude Opus 5.5 (xhigh) moved from #8 to #10 in the BenchLeader Index. (details)
• GPT-6.1 Sol (max) moved from #10 to #11 in the BenchLeader Index. (details)

Price changes

• GPT-5.6 Sol Pro (max) input price moved from $4.00/M to $2.00/M (50% down). (details)
• GPT-5.6 Sol Pro (max) output price moved from $20.00/M to $10.00/M (50% down). (details)
• GPT-5.6 Sol Pro (xhigh) input price moved from $4.00/M to $2.00/M (50% down). (details)
• GPT-5.6 Sol Pro (xhigh) output price moved from $20.00/M to $10.00/M (50% down). (details)
• GPT-5.6 Sol Pro input price moved from $4.00/M to $2.00/M (50% down). (details)
• GPT-5.6 Sol Pro output price moved from $20.00/M to $10.00/M (50% down). (details)

New benchmark results

• SciCode: 18 new results, including Claude Opus 5.5, Gemini 4 Argon, Claude Opus 5.5… (board)
• APEX-Agents: 11 new results, including Claude Opus 5.5, Qwen3.8 Max, MiMo-V2.6-Pro… (board)
• CursorBench: 9 new results, including Claude Opus 5.5, Claude Opus 5.5, Claude Opus 5.5… (board)
• ProofBench: 5 new results, including Muse Spark 1.3, MiMo-V2.6-Pro, Step 5 Preview… (board)
• GSO-Bench: 4 new results, including GPT-6 Astra, Claude Fable 5, Claude Fable 5.1… (board)
• GDP.pdf: 3 new results, including GPT-6 Astra, DeepSeek V4.1 Flash, Grok 4.7 (board)
• ALE-Bench: 2 new results, including Claude Opus 5.5, MiMo-V2.6-Pro (board)
• FrontierSWE: 2 new results, including Gemini 4 Argon, GLM 5.3 Flash (board)
• Vending-Bench 2: 2 new results, including Claude Opus 5.5, Grok 4.7 (board)
• Blueprint-Bench 2: 2 new results, including Claude Opus 5.5, Grok 4.7 (board)
• WeirdML: 1 new result, including Gemini 3.8 Flash (board)
• LMCA: 1 new result, including Grok 4.7 (board)

Score revisions

• Claude Sonnet 5 (high) on SciCode: 48.6% → 54.3%. (details)
• Qwen3.8 Max (xhigh) on FrontierSWE: 15.8% → 17.8%. (details)

Speed changes

• GPT-5.5 Pro output speed changed from 5 tok/s to 9 tok/s. (details)
• Muse Spark 1.1 time to first token changed from 2.02 s to 3.55 s. (details)
• GPT-5.4 Pro output speed changed from 2 tok/s to 4 tok/s. (details)
• Claude Sonnet 5 output speed changed from 52 tok/s to 75 tok/s. (details)
• GPT-5 nano time to first token changed from 1.54 s to 2.35 s. (details)
• Ling 3.0 Flash VL output speed changed from 48 tok/s to 80 tok/s. (details)
• Qwen3.5-35B-A3B time to first token changed from 0.76 s to 1.37 s. (details)
• Qwen3.5-27B output speed changed from 16 tok/s to 28 tok/s. (details)
• Qwen3.5 Flash output speed changed from 49 tok/s to 74 tok/s. (details)
• GLM 4.7 Flash output speed changed from 25 tok/s to 44 tok/s. (details)
• GLM 4.5V output speed changed from 22 tok/s to 31 tok/s. (details)
• Qwen3.5-9B time to first token changed from 0.74 s to 1.04 s. (details)
• Gemini 2.5 Flash-Lite output speed changed from 81 tok/s to 137 tok/s. (details)
• MiniMax-M1 output speed changed from 21 tok/s to 32 tok/s. (details)
• Qwen3.6 35B A3B output speed changed from 47 tok/s to 78 tok/s. (details)
• …and 16 more.

Today's top five

1. Claude Opus 5.5 — 72.0
2. Claude Fable 5.1 — 70.8
3. GPT-6 Astra — 70.2
4. Claude Opus 5.5 — 70.2
5. Gemini 4 Argon — 70.0

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.