Mistral Large 4.0.8 added, GPT-6 Astra up to #8 (30 Oct)

BenchLeader refreshed 1675 models from 26 of 26 data feeds (benchmark publishers, aggregators and price lists). 1 new model, 2 top-10 rank changes, 94 price changes, 59 new or updated benchmark results across 26 data feeds.

New models

Mistral Large 4.0.8 (Mistral AI) is now listed at $1.05/M blended; no independent results yet. (details)

Movement in the top 10

GPT-6 Astra (xhigh) moved from #9 to #8 in the BenchLeader Index. (details)
Claude Fable 5 moved from #8 to #9 in the BenchLeader Index. (details)

Price changes

Qwen3.7 (max) cached input price moved from $0.50/M to $0.33/M (34% down). (details)
Qwen3.6 Plus cached input price moved from $0.050/M to $0.028/M (45% down). (details)
Qwen3.8 Max (max) cached input price moved from $0.25/M to $0.22/M (11% down). (details)
Qwen3.7 Plus cached input price moved from $0.080/M to $0.055/M (31% down). (details)
Qwen3.5 Plus (thinking) input price moved from $0.40/M to $0.12/M (71% down). (details)
Qwen3.5 Plus (thinking) output price moved from $2.40/M to $0.69/M (71% down). (details)
Mistral Large 4 input price moved from $1.36/M to $0.70/M (49% down). (details)
Mistral Large 4 output price moved from $4.18/M to $2.10/M (50% down). (details)
Mistral Large 4 cost per task moved from $5.54/M to $4.21/M (24% down). (details)
Mistral Large 4 cached input price moved from $0.14/M to $0.070/M (50% down). (details)
Qwen3.5 Flash input price moved from $0.10/M to $0.029/M (71% down). (details)
Qwen3.5 Flash output price moved from $0.40/M to $0.29/M (28% down). (details)
Qwen3 32B input price moved from $0.70/M to $0.29/M (59% down). (details)
Qwen3 32B output price moved from $2.80/M to $1.15/M (59% down). (details)
Qwen3 235B A22B input price moved from $0.70/M to $0.29/M (59% down). (details)
Qwen3 235B A22B output price moved from $2.80/M to $1.15/M (59% down). (details)
Mistral Large 4 (high) cached input price moved from $0.14/M to $0.070/M (50% down). (details)
Qwen-Plus input price moved from $0.40/M to $0.12/M (71% down). (details)
Qwen-Plus output price moved from $1.20/M to $0.29/M (76% down). (details)
Qwen-Plus cached input price moved from $0.080/M to $0.023/M (71% down). (details)
Qwen3 235B A22B (no reasoning) input price moved from $0.70/M to $0.29/M (59% down). (details)
Qwen3 235B A22B (no reasoning) output price moved from $2.80/M to $1.15/M (59% down). (details)
Qwen3 14B input price moved from $0.35/M to $0.14/M (59% down). (details)
Qwen3 14B output price moved from $1.40/M to $0.57/M (59% down). (details)
Qwen3.6 Flash input price moved from $0.25/M to $0.17/M (34% down). (details)
Qwen3.6 Flash output price moved from $1.50/M to $0.99/M (34% down). (details)
Qwen3.6 Flash cached input price moved from $0.025/M to $0.017/M (34% down). (details)
DeepSeek V4 Pro 0424 (high) input price moved from $0.44/M to $1.67/M (280% up). (details)
DeepSeek V4 Pro 0424 (high) output price moved from $0.87/M to $3.34/M (284% up). (details)
DeepSeek V4 Pro 0424 (max) input price moved from $0.44/M to $1.41/M (220% up). (details)
…and 64 more.

New benchmark results

Arena Agent: 6 new results, including GPT-6.1 Terra, Muse Wonder Full, DeepSeek V4.1 Flash 1017… (board)
AA Intelligence Index v5.0: 6 new results, including GLM 5.3 Flash, Kimi K3, Qwen3.8 27B… (board)
FrontierCode 2.0 (AA): 6 new results, including GLM 5.3 Flash, Kimi K3, Qwen3.8 27B… (board)
AA-Omniscience v1.1: 4 new results, including Gemma 4 31B, Granite 4.2 30B, Granite 4.2 3B… (board)
AA-Omniscience v1.1: accuracy: 4 new results, including Gemma 4 31B, Granite 4.2 30B, Granite 4.2 3B… (board)
AA-Omniscience v1.1: non-hallucination: 4 new results, including Gemma 4 31B, Granite 4.2 30B, Granite 4.2 3B… (board)
Terminal-Bench-Science 0.1 (AA): 1 new result, including Mistral Medium 3.5 (board)
HLE-Diamond (AA): 1 new result, including Granite 4.2 30B (board)
CritPt v2 hard subset (AA): 1 new result, including Granite 4.2 3B (board)
AA Progress Index v1: 1 new result, including Mistral Large 4 (board)

Score revisions

Gemini 4 Carbon (high) on AA Progress Index v1: 351.6 → 350.8. (details)
Kimi K3.1 Pro on GDPval-AA v2.1: 1846 → 1858. (details)
DeepSeek V4.1 Flash 1017 (max) on OpenRouter token share: 21.1% → 22.3%. (details)
GPT-6 Luna (max) on OpenRouter token share: 5.5% → 6.3%. (details)
Mistral Large 4 on LiveBench 2.0: 43.9% → 63.5%. (details)
Mistral Large 4 on LiveBench Long Agentic: 8.8% → 27.5%. (details)
Mistral Large 4 on LiveBench 2.0 Reasoning: 83.9% → 85.2%. (details)
Mistral Large 4 on LiveBench 2.0 Coding: 60% → 95.4%. (details)
Mistral Large 4 on LiveBench Browser: 0.6% → 17.4%. (details)
Mistral Large 4 on LiveBench Research: 28.1% → 79.7%. (details)
Mistral Large 4 on LiveBench Design: 65.6% → 76.3%. (details)
GPT-6 Luna (xhigh) on OpenRouter token share: 5.5% → 6.3%. (details)
MiniMax M3.1 Flash on OpenRouter token share: 3.5% → 2.2%. (details)
GPT-6 Luna (high) on OpenRouter token share: 5.5% → 6.3%. (details)
GPT-6 Luna (medium) on OpenRouter token share: 5.5% → 6.3%. (details)
GPT-6 Luna (low) on OpenRouter token share: 5.5% → 6.3%. (details)
GPT-6 Luna (no reasoning) on OpenRouter token share: 5.5% → 6.3%. (details)
DeepSeek V4.1 Flash 1017 on OpenRouter token share: 21.1% → 22.3%. (details)
GPT-6 Luna on OpenRouter token share: 5.5% → 6.3%. (details)
DeepSeek V4.1 Flash 1017 (no reasoning) on OpenRouter token share: 21.1% → 22.3%. (details)
…and 5 more.

Speed changes

GPT-6 Astra time to first token changed from 1.48 s to 3.34 s. (details)
Claude Opus 5 time to first token changed from 1.88 s to 2.80 s. (details)
GPT-6.1 Terra time to first token changed from 3.51 s to 2.26 s. (details)
GPT-5.6 Sol time to first answer changed from 20.30 s to 29.18 s. (details)
Muse Wonder output speed changed from 53 tok/s to 81 tok/s. (details)
Claude Fable 5.1 time to first answer changed from 2.35 s to 3.90 s. (details)
GPT-5.5 time to first answer changed from 3.26 s to 7.89 s. (details)
GPT-5.5 response time changed from 10.61 s to 15.24 s. (details)
GPT-5.6 Terra time to first answer changed from 9.24 s to 14.47 s. (details)
GPT-5.6 Terra response time changed from 13.94 s to 18.99 s. (details)
GPT-5.2 time to first answer changed from 65.42 s to 99.55 s. (details)
GPT-5.2 response time changed from 72.43 s to 106.65 s. (details)
GPT-5.2 time to first token changed from 3.92 s to 2.42 s. (details)
Claude Sonnet 5 time to first answer changed from 2.49 s to 4.63 s. (details)
Kimi K2.5 time to first token changed from 1.61 s to 0.97 s. (details)
…and 59 more.

Today's top five

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.