Enter opens the highlighted result. Esc closes.

Best open-weight LLM for math

DeepSeek V4.1 Flash 1017 leads
85.3
models ranked
43
downloads of #1, 30 days
51.9K
lead over #2
+1.1

Math Index · read October 22

  1. New MiniMax M3.1 Flash October 19
  2. New DeepSeek V4.1 Flash 1017 October 17
  3. New E Glimpse October 17
  4. New Mercury 3 October 16

As of October 22, 2026, DeepSeek V4.1 Flash 1017 leads 43 open-weight models on Math Index with 85.3, ahead of MoonshotAI Kimi K3 (84.2) and NVIDIA Nemotron 3 Ultra (83.2).

The full best LLM for math ranking combines this benchmark with others.

Top 20 of 43, higher is better

Embed this data
Best open-weight LLM for math
Rank Model Math Index Coding evidence Weights Coding index Output $/Mtok
1
DeepSeek V4.1 Flash 1017 DeepSeek
85.3
 deepseek-ai/DeepSeek-V4.1-Flash-1017 70.3 $0.60
2
Kimi K3 Moonshot AI
84.2
 moonshotai/Kimi-K3 62.3 $15.00
3
Nemotron 3 Ultra NVIDIA
83.2
 nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 30.0 $2.40
4
DeepSeek V4.1 Flash DeepSeek
82.2
 deepseek-ai/DeepSeek-V4.1-Flash 62.5 $0.84
5
Gemma 4 26B A4B Google
80.9
 google/gemma-4-26B-A4B-it Not enough data $0.60
6
gpt-oss-120b OpenAI
80.9
 openai/gpt-oss-120b Not enough data $0.50
7
Qwen3 235B A22B Thinking 2507 Qwen
80.7
 Qwen/Qwen3-235B-A22B-Thinking-2507 Not benchmarked $3.50
8
GLM 5.3 Z.ai
80.5
 zai-org/GLM-5.3 70.6 $4.40
9
DeepSeek V4 Pro 0813 DeepSeek
80.0
 deepseek-ai/DeepSeek-V4-Pro-0813 56.1 $1.98
10
GLM 4.7 Z.ai
79.1
 zai-org/GLM-4.7 55.1 $2.20
11
GLM 5 Z.ai
77.8
 zai-org/GLM-5 41.7 $3.20
12
Kimi K2.6 Moonshot AI
77.1
 moonshotai/Kimi-K2.6 52.1 $4.00
13
GLM 5.3 Flash Z.ai
76.6
 zai-org/GLM-5.3-Flash 57.7 $0.50
14
DeepSeek V4 Flash 0731 DeepSeek
76.0
 deepseek-ai/DeepSeek-V4-Flash-0731 51.8 $0.28
15
Kimi K2.7 Code Moonshot AI
74.8
 moonshotai/Kimi-K2.7-Code 53.4 $4.00
16
Gemma 4 31B Google
74.6
 google/gemma-4-31B-it 39.0 $0.40
17
GLM 5.2 Z.ai
74.3
 zai-org/GLM-5.2 61.1 $4.40
18
DeepSeek V4 Pro 0423 DeepSeek
71.5
 deepseek-ai/DeepSeek-V4-Pro 38.7 $2.716
19
MiniMax M3 MiniMax
71.3
 MiniMaxAI/Minimax-M3 39.2 $1.20
20
T Inkling Small Thinking Machines
69.8
 thinkingmachines/Inkling-Small 27.3 $1.20

How this list is built

Generated from the dataset on every refresh. Nobody picks the order by hand.

Models on the Math Index board that OpenRouter links to a public Hugging Face repository, by their Math Index result. We do not check the license; read it on the repository page.

Other lists

Rankings beyond coding · Methodology