Gemini 4 Hydrogen added, GPT-6 Astra up to #6 (12 Nov 2026)
Gemini 4 Hydrogen joins the board; GPT-6 Astra moved #7 → #6. 1 new model, 6 top-10 rank changes, 17 price changes, 151 new benchmark results.
BenchLeader refreshed 1700 models from 29 of 29 data feeds (benchmark publishers, aggregators and price lists). 1 new model, 6 top-10 rank changes, 17 price changes, 151 new or updated benchmark results across 29 data feeds.
New models
- Gemini 4 Hydrogen (Google) appeared with 1 benchmark result. (details)
Movement in the top 10
- GPT-6 Astra (max) moved from #7 to #6 in the BenchLeader Index. (details)
- GPT-6 Astra moved from #8 to #7 in the BenchLeader Index. (details)
- Claude Fable 5.1 (max) moved from #9 to #8 in the BenchLeader Index. (details)
- GPT-6.1 Sol (max) moved from #10 to #9 in the BenchLeader Index. (details)
- Claude Fable 5.1 moved from #11 to #10 in the BenchLeader Index. (details)
- Claude Fable 5.2 moved from #6 to #15 in the BenchLeader Index. (details)
Price changes
- Gemini 4 Carbon (high) input price moved from $4.00/M to $2.00/M (50% down). (details)
- Gemini 4 Carbon (high) output price moved from $20.00/M to $10.00/M (50% down). (details)
- Gemini 4 Carbon (high) cost per task moved from $7.76/M to $5.48/M (29% down). (details)
- Gemini 4 Carbon (medium) input price moved from $4.00/M to $2.00/M (50% down). (details)
- Gemini 4 Carbon (medium) output price moved from $20.00/M to $10.00/M (50% down). (details)
- Gemini 4 Carbon (medium) cost per task moved from $6.59/M to $4.68/M (29% down). (details)
- Gemini 4 Argon (high) input price moved from $4.00/M to $2.00/M (50% down). (details)
- Gemini 4 Argon (high) output price moved from $20.00/M to $10.00/M (50% down). (details)
- Gemini 4 Argon (high) cost per task moved from $9.95/M to $4.97/M (50% down). (details)
- Gemini 4 Carbon (low) input price moved from $4.00/M to $2.00/M (50% down). (details)
- Gemini 4 Carbon (low) output price moved from $20.00/M to $10.00/M (50% down). (details)
- Gemini 4 Carbon (low) cost per task moved from $7.19/M to $5.22/M (27% down). (details)
- Hy4 cached input price moved from $0.021/M to $0.025/M (19% up). (details)
- Gemma 4 26B A4B (thinking) input price moved from $0.10/M to $0.090/M (10% down). (details)
- Gemma 4 26B A4B (thinking) output price moved from $0.34/M to $0.32/M (6% down). (details)
- Gemma 4 31B (no reasoning) input price moved from $0.16/M to $0.15/M (6% down). (details)
- Gemma 4 26B A4B (no reasoning) input price moved from $0.12/M to $0.13/M (8% up). (details)
New benchmark results
- FACTS Search: 10 new results, including Gemini 3 Flash, Grok 4.6, GPT-5.6 Sol… (board)
- PostTrainBench: 8 new results, including Claude Opus 5.5, GPT-6 Astra, Claude Fable 5.1… (board)
- AIME 2026: 7 new results, including Gemini 3.5 Flash, Kimi K3, Gemini 3.1 Pro… (board)
- HMMT February 2026: 7 new results, including Gemini 3.5 Flash, Kimi K3, Gemini 3.1 Pro… (board)
- MathArena Apex: 7 new results, including Gemini 3.5 Flash, Kimi K3, Gemini 3.1 Pro… (board)
- FrontierCode 2.0 (AA): 3 new results, including Qwen3.8-Flash-Next, Solar Pro 4, A.X-K2 (board)
- Arena Agent: 2 new results, including Claude Fable 5.2, Muse Ember (board)
- PRBench Finance: 2 new results, including Claude Fable 5.2, Kimi K3.1 Pro (board)
- PRBench Legal: 2 new results, including Claude Fable 5.2, Kimi K3.1 Pro (board)
- FORTRESS: 2 new results, including Kimi K3.1 Pro, Gemini 4 Carbon (board)
- AA Intelligence Index v5.0.1: 2 new results, including Qwen3.8-Flash-Next, Solar Pro 4 (board)
- Terminal-Bench-Science 0.1 (AA): 2 new results, including Qwen3.8-Flash-Next, A.X-K2 (board)
Score revisions
- Claude Opus 5.5 (max) on Arena Agent: 20 → 19.5. (details)
- GPT-6 Astra (max) on Design Arena Website: 1330 → 1339. (details)
- GPT-6 Astra (max) on Design Arena Agentic: 1385 → 1391. (details)
- Claude Fable 5.1 (max) on Arena Agent: 14.3 → 14.8. (details)
- GPT-6 Astra (xhigh) on Design Arena Website: 1331 → 1341. (details)
- Claude Fable 5.2 on Epoch Capabilities Index: 174.2 → 173.4. (details)
- Claude Fable 5.2 on Design Arena Website: 1320 → 1353. (details)
- Claude Fable 5.2 on Design Arena Agentic: 1362 → 1355. (details)
- GPT-5.6 Sol (max) on PostTrainBench: 36.2% → 31%. (details)
- Claude Fable 5 (max) on PostTrainBench: 41.8% → 37.3%. (details)
- Claude Opus 5 on PostTrainBench: 35% → 35.7%. (details)
- Gemini 4 Argon on Design Arena Agentic: 1264 → 1269. (details)
- GPT-5.5 (xhigh) on PostTrainBench: 27.2% → 23.9%. (details)
- GPT-6.1 Sol on Design Arena Website: 1309 → 1315. (details)
- Muse Spark 1.3 (max) on Design Arena Website: 1342 → 1353. (details)
- Muse Wonder (max) on Design Arena Website: 1368 → 1379. (details)
- Kimi K3.1 Pro on Epoch Capabilities Index: 162 → 161.4. (details)
- Kimi K3.1 Pro on Design Arena Agentic: 1282 → 1287. (details)
- Claude Opus 4.8 (high) on PostTrainBench: 33.8% → 30.7%. (details)
- Muse Spark 1.3 on LiveBench 2.0 Coding: 83.1% → 84.4%. (details)
- …and 66 more.
Speed changes
- Claude Opus 5 time to first answer changed from 8.95 s to 13.73 s. (details)
- GPT-5.6 Sol time to first answer changed from 9.58 s to 12.96 s. (details)
- GPT-5.4 Pro time to first token changed from 16.10 s to 8.70 s. (details)
- Muse Spark 1.3 time to first token changed from 2.12 s to 3.19 s. (details)
- GLM-5 time to first answer changed from 22.75 s to 31.40 s. (details)
- GLM-5 thinking time changed from 20.01 s to 28.64 s. (details)
- Muse Spark 1.3 output speed changed from 186 tok/s to 97 tok/s. (details)
- GPT-5.6 Terra time to first answer changed from 12.29 s to 20.72 s. (details)
- GPT-5.6 Terra response time changed from 40.75 s to 62.62 s. (details)
- GLM 5.2 time to first token changed from 1.31 s to 0.64 s. (details)
- Qwen3.8 Max time to first token changed from 1.91 s to 2.87 s. (details)
- Grok 4.20 Multi-Agent time to first token changed from 23.15 s to 11.81 s. (details)
- Claude Sonnet 4.6 time to first token changed from 1.21 s to 0.76 s. (details)
- GPT-6 Sol time to first answer changed from 5.86 s to 8.12 s. (details)
- Inkling 2 Small time to first answer changed from 46.89 s to 63.89 s. (details)
- …and 109 more.
Source status
- Source t3code is updating again.
Today's top five
- Claude Fable 5.2 (max) — 72.7
- Gemini 4 Carbon (high) — 68.9
- Claude Opus 5.5 (max) — 68.4
- GPT-6 Astra (max) — 67.3
- 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.