Value AI model rankings
The Value track ranks 39 AI models by AMR aggregate score per blended dollar (25% input, 75% output pricing). Snapshot 2026-10-14, updated weekly.
Snapshot 2026-10-14 · updated weekly
| Rank | Model | Vendor | Score / $ | GPQA | Context | Price (in / out) | Speed |
|---|---|---|---|---|---|---|---|
1— | Qwen3.8 Flash | Alibaba | 367.9 | 66.8% | 256K | $0.06 / $0.25 | 412 |
2NEW | Laya 1 | Laya Research | 340.9 | 38.4% | 64K | $0.05 / $0.20 | 368 |
3NEW | Jev 1 | Jev Labs | 281.2 | 59.4% | 128K | $0.08 / $0.30 | 322 |
4— | MiniMax H3 Flash | MiniMax | 242.5 | 72.6% | 512K | $0.10 / $0.40 | 358 |
5NEW | Clef 2 | Clef AI | 233.7 | 56.9% | 128K | $0.10 / $0.35 | 276 |
6▲3 | GLM-5.3 Flash | Zhipu AI | 207.7 | 80.4% | 256K | $0.12 / $0.50 | 268 |
7NEW | Strata 2 | Strata Labs | 191.8 | 62.8% | 128K | $0.12 / $0.45 | 248 |
8— | Mercury 2.5 | Inception Labs | 190.8 | 48.2% | 128K | $0.10 / $0.40 | 583 |
9— | Kimi K3 Flash | Moonshot AI | 169.8 | 70.2% | 256K | $0.14 / $0.55 | 296 |
10▲2 | Gemini 3.8 Flash | 169.6 | 78.9% | 1M | $0.15 / $0.60 | 342 | |
11— | Qwen3.8 27B | Alibaba | 163.7 | 76.8% | 256K | $0.15 / $0.60 | 186 |
12— | Gemma 4 27B | 126.3 | 67.4% | 256K | $0.18 / $0.70 | 174 | |
13— | MiMo-V2.6-Pro | Xiaomi | 96.0 | 75.4% | 256K | $0.25 / $1 | 142 |
14— | Celeris-1 | Celeris | 89.5 | 42.6% | 128K | $0.20 / $0.80 | 1431 |
15— | DeepSeek V4.5 | DeepSeek | 86.7 | 83.7% | 256K | $0.30 / $1.20 | 112 |
16NEW | DeepSeek Harness | DeepSeek | 78.1 | 86.5% | 256K | $0.35 / $1.40 | 105 |
17— | GLM-5.3 | Zhipu AI | 69.8 | 85.2% | 256K | $0.40 / $1.50 | 118 |
18— | Muse Wonder | Meta | 64.9 | 85.9% | 256K | $0.45 / $1.60 | 126 |
19▲1 | Qwen3.8 Next | Alibaba | 56.7 | 82.9% | 512K | $0.50 / $1.80 | 104 |
20— | Kimi K3 | Moonshot AI | 53.8 | 93.5% | 512K | $0.50 / $2 | 98 |
21▲1 | MiniMax H3 | MiniMax | 49.0 | 85.8% | 1M | $0.60 / $2.20 | 121 |
22— | Mistral Large 3.5 | Mistral AI | 41.2 | 80.7% | 256K | $0.80 / $2.40 | 132 |
23— | Muse Wonder Full | Meta | 33.1 | 88.1% | 512K | $0.90 / $3.20 | 88 |
24— | Grok-4 Fast Reasoning | xAI | 31.9 | 81.5% | 2M | $0.80 / $3.20 | 210 |
25NEW | Claude Haiku 5.5 | Anthropic | 27.5 | 83.2% | 500K | $0.80 / $4 | 232 |
26— | Gemini 3.1 Pro | 19.7 | 79.8% | 1M | $1.25 / $5 | 172 | |
27— | GPT-5.6 Luna | OpenAI | 17.7 | 74.2% | 512K | $1.20 / $5.50 | 286 |
28— | Gemini 3 Pro | 16.6 | 80.9% | 1M | $1.50 / $6 | 168 | |
29— | Gemini 4.0 | 11.1 | 87.9% | 2M | $2.20 / $10 | 158 | |
30▲2 | Gemini 4 Argon | 9.7 | 92.1% | 2M | $2.50 / $12 | 165 | |
31— | Claude Sonnet 5.5 | Anthropic | 7.9 | 92.8% | 500K | $3 / $15 | 156 |
32— | GPT-5.6 Sol | OpenAI | 7.1 | 88.4% | 512K | $2.80 / $16 | 134 |
33— | Claude Fable 5 | Anthropic | 6.5 | 91.9% | 400K | $3.50 / $18 | 138 |
34— | Claude Fable 5.1 | Anthropic | 5.9 | 93.4% | 400K | $4 / $20 | 142 |
35— | GPT-6 Sol | OpenAI | 4.8 | 90.2% | 800K | $4 / $24 | 118 |
36▼1 | GPT-6 Astra | OpenAI | 4.0 | 96% | 1M | $5 / $30 | 128 |
37▼1 | Claude Opus 5 | Anthropic | 3.9 | 90.7% | 400K | $6 / $30 | 96 |
38▲1 | GPT-6.1 Sol | OpenAI | 3.4 | 94.9% | 1M | $6 / $36 | 112 |
39— | Claude Opus 5.5 | Anthropic | 3.0 | 95.8% | 500K | $8 / $40 | 88 |
Value FAQ
Which AI model is best for Value?
Qwen3.8 Flash leads the Value track as of October 2026. The full table above ranks all 39 tracked models on this dimension, with the underlying values shown for every row.
How is the Value ranking calculated?
This track ranks models by AMR aggregate score per blended dollar (25% input, 75% output pricing). Scores are re-verified weekly against public benchmark sources and API rate cards, and the snapshot date is shown on every table so you can see exactly how fresh each value is.