State of the leaderboard: week 42, 12–18 October 2026

Complete week: data as of 18 Oct 2026, from 7 daily updates

The week in 6 sentences

In week 42 of 2026 (12–18 October 2026), Claude Opus 5.5 (Anthropic) held the #1 spot on the BenchLeader Index at 69.2, 0.3 points ahead of Gemini 4 Argon (68.9).
1 model entered the top 10 (Muse Spark at #10), pushing out Muse Spark 1.3; 4 other models changed places.
16 new models arrived; the highest-ranked is Muse Wonder (Meta) at #14 with an index of 63.9.
The biggest riser was Gemini 4 Argon, up 2.3 points to 68.9; the biggest faller Devstral Medium, down 3.7 to 40.0.
28 models cut a list price and 19 models raised one; the largest cut was Qwen3 14B's output price, 94% lower.
1,400 new benchmark results were published, the most on AutomationBench (162).

Top 10 at the end of the week

The BenchLeader Index, one row per model at its best reasoning effort, against the table at the end of the previous week. Left the top 10: Muse Spark 1.3.

#	Model	Provider	Index	Last week	Move
1	Claude Opus 5.5	Anthropic	69.2	1	–
2	Gemini 4 Argon	Google	68.9	7	↑5
3	GPT-6 Astra	OpenAI	68.1	3	–
4	Claude Fable 5.1	Anthropic	67.8	2	↓2
5	GPT-6.1 Sol	OpenAI	67.6	5	–
6	Claude Fable 5	Anthropic	67.3	4	↓2
7	Claude Opus 5	Anthropic	66.5	6	↓1
8	GPT-5.6 Sol	OpenAI	65.7	8	–
9	Claude Sonnet 5.5	Anthropic	64.9	9	–
10	Muse Spark	Meta	64.7	–	new

New models and their first results

Muse Wonder (Meta): #14 on the index at 63.9. First results: 31.4% on EBR-bench; 10.0% on Harvey LAB-AA v1.1; 99.4% on OTIS Mock AIME; 60.3% on SciCode.
Muse Wonder Full (Meta): not enough independent results for an index yet. First results: 13.1% on Harvey LAB-AA v1.1; 79.8% on IOI; 87.4% on LegalBench; 64.6 on Vals Index.
musespark-1-3: not enough independent results for an index yet
lfm-2-5-24b-a2b: not enough independent results for an index yet
Nemotron 3.5 Super VL 120B A12B (NVIDIA): not enough independent results for an index yet
Muse Spark Open (Meta): not enough independent results for an index yet. First results: 46.6 on Vals Index; 21.2% on Code Migration; 53.7% on Excel Modeling Benchmark; 61.3% on Finance Agent v2.
Nemotron 3.5 Bolt 74B-A9B (NVIDIA): not enough independent results for an index yet. First results: 15.5% on AutomationBench; 2.8% on GDP.pdf (AA).
Nemotron 3.5 Super VL (NVIDIA): not enough independent results for an index yet
GLM-Legal (Zhipu AI): not enough independent results for an index yet
Muse Spark Open 816B A42B (Meta): #81 on the index at 57.4
Beam 1.0 (Reflection): #115 on the index at 54.1
nemotron-3-5-super-vl-120b-a12b-nvfp4: not enough independent results for an index yet
Beam 1.0 501B A23B (Reflection): not enough independent results for an index yet
Seed2.1 Pro (ByteDance): not enough independent results for an index yet
DeepSeek V4.1 Flash 1017 (DeepSeek): not enough independent results for an index yet
Salesforce Launa Omni 1 (Salesforce): not enough independent results for an index yet

Biggest movers on the index

Model	Provider	12 Oct 2026	18 Oct 2026	Change
Devstral Medium	Mistral AI	43.7	40.0	-3.7
Gemini 4 Argon	Google	66.6	68.9	+2.3
DeepSeek V4 Pro	DeepSeek	55.7	57.5	+1.8
Mimo v2 Omni	Xiaomi	57.4	58.7	+1.3
Mistral Large 4	Mistral AI	56.4	57.5	+1.1
Claude Fable 5.1	Anthropic	68.7	67.8	-0.9
GPT-4	OpenAI	45.0	44.1	-0.9
Grok 4.7	SpaceXAI	58.6	59.4	+0.8
Kimi K3	Moonshot AI	63.7	63.0	-0.7
DeepSeek V4 Flash	DeepSeek	58.6	59.3	+0.7

Price changes

Cheaper (28)

Qwen3 14B (thinking) output price moved from $4.20/M to $0.24/M (94% down).
Qwen3 Coder 480B input price moved from $1.50/M to $0.30/M (80% down).
Qwen3 30B A3B (thinking) output price moved from $2.40/M to $0.55/M (77% down).
Qwen3 Max (max) output price moved from $6.00/M to $1.43/M (76% down).
Qwen3 8B (no reasoning) output price moved from $0.70/M to $0.20/M (71% down).
Qwen3 Coder 30B A3B Instruct input price moved from $0.45/M to $0.15/M (67% down).
Qwen3 235B A22B (thinking) output price moved from $8.40/M to $2.80/M (67% down).
Grok 4.20 (thinking) output price moved from $6.00/M to $2.50/M (58% down).
QwQ-32B input price moved from $0.66/M to $0.29/M (57% down).
DeepSeek V4 Pro (high) output price moved from $1.98/M to $0.87/M (56% down).
Qwen3 30B A3B 2507 output price moved from $0.80/M to $0.40/M (50% down).
Qwen3 VL 235B A22B Instruct (thinking) input price moved from $0.40/M to $0.22/M (45% down).
…and 16 more.

More expensive (19)

DeepSeek V4 Pro (no reasoning) output price moved from $0.87/M to $3.96/M (355% up).
Llama 3.1 8B input price moved from $0.020/M to $0.040/M (100% up).
GPT-5.6 Sol Pro (max) input price moved from $2.00/M to $4.00/M (100% up).
Qwen3 VL 30B A3B Instruct (thinking) input price moved from $0.20/M to $0.35/M (75% up).
Qwen3 32B (thinking) input price moved from $0.16/M to $0.24/M (50% up).
Nemotron 3 Super 120B A12b input price moved from $0.20/M to $0.30/M (50% up).
Qwen3 Coder Next input price moved from $0.35/M to $0.50/M (43% up).
Qwen3 VL 8B Instruct (thinking) input price moved from $0.18/M to $0.25/M (39% up).
DeepSeek V4.1 Flash 1017 input price moved from $0.15/M to $0.20/M (33% up).
DeepSeek V4 Flash (max) input price moved from $0.11/M to $0.14/M (27% up).
gpt-oss-20b (high) output price moved from $0.18/M to $0.22/M (22% up).
gpt-oss-120b (low) output price moved from $0.50/M to $0.59/M (18% up).
…and 7 more.

List prices per million tokens; each model's full price history is on its page and the LLM API pricing tracker.

Other weeks

← Week 43, 19–25 Oct 2026 · Week 41, 5–11 Oct 2026 → · The weekly review on the blog

Benchmark categories: AI agent benchmarks · AI coding benchmarks · Reasoning and math benchmarks · Multimodal AI benchmarks · New AI model releases

How this is built: the BenchLeader Index is recomputed with every update from independent benchmark publishers (see the methodology), and a maker's own launch numbers are never used. Past ranks are rebuilt from the daily index history of the models listed today.

Cite as: BenchLeader, “AI leaderboard week 42, 12–18 Oct 2026: who was #1”, data as of 18 Oct 2026, https://www.benchleader.com/state/2026-w42.