Home/Overall leaderboard

Overall leaderboard

Total 30 models. Sorted by v5 intelligence index by default; models without a v5 score follow, showing their v4.3.2 score in grey. Click any column header to re-sort.

# Model / vendor v5 v4.3.2 GPQA TB 4.0 LBench D Input Output tok/s Context Type
1 Gemini 4 CarbonGoogle 60.0 — — 65.2 — $4.00 $20.00 88 1M Closed
2 Claude Opus 5.5Anthropic 59.1 57.6 93.9 59.6 86.9 $4.00 $20.00 94 1M Closed
3 GPT-6 AstraOpenAI 56.9 — 95.8 59.1 82.0 $10.00 $50.00 81 1.05M Closed
4 GPT-6.1 SolOpenAI 54.6 — 95.4 56.1 82.1 $2.00 $10.00 77 1.05M Closed
5 Claude Sonnet 5.5Anthropic 54.5 56.0 95.6 63.6 85.0 $2.00 $10.00 135 1M Closed
6 GPT-6.1 TerraOpenAI 51.3 — 94.3 55.6 82.3 $1.00 $5.00 94 1.05M Closed
7 Gemini 4 ArgonGoogle 50.4 — 96.4 57.1 73.7 $2.00 $10.00 80 1M Closed
8 Muse Wonder FullMeta 50.0 56.0 — — — — — — 1M Closed
9 Claude Fable 5.1Anthropic 49.0 53.0 — 52.0 83.4 $10.00 $50.00 57 1M Closed
10 Kimi K3.1 Pro Open weightsMoonshot AI 46.0 55.1 — 67.2 77.8 $2.00 $10.00 28 1M Open weights
11 Muse WonderMeta 44.0 — — — — $4.00 $20.00 143 1M Closed
12 Claude Haiku 5.5Anthropic 36.0 43.0 — — 83.1 $0.10 $0.50 235 1M Closed
13 Grok 4.7xAI 34.3 46.4 92.7 25.8 78.0 $2.00 $6.00 105 500K Closed
14 Qwen3.8 MaxAlibaba Qwen 34.0 — 92.7 38.9 78.6 $2.00 $6.00 88 1M Closed
15 DeepSeek V4.1 Flash Open weightsDeepSeek 33.0 45.1 89.8 19.7 83.4 $0.30 $1.20 148 1.05M Open weights
16 GLM-5.3 Open weightsZ.ai 32.0 44.8 90.9 41.9 — $1.40 $4.40 77 1M Open weights
17 MiMo-V2.6 Pro Open weightsXiaomi 30.0 46.3 59.5 31.3 78.5 $0.35 $1.40 81 1M Open weights
18 MiniMax M3.1 FlashMiniMax 29.0 — — — 74.0 $0.15 $0.60 118 1M Closed
19 Mistral Large 4 Open weightsMistral 26.0 — 87.6 — — $0.70 $2.10 — 1M Open weights
20 DeepSeek V4 Pro Open weightsDeepSeek 23.0 — — 14.1 — $1.32 $3.96 72 1.05M Open weights
21 Kimi K3 Open weightsMoonshot AI — 43.6 93.1 17.2 77.2 $3.00 $15.00 30 1M Open weights
22 GLM-5.3 Flash Open weightsZ.ai — 41.8 — — 77.2 $0.15 $0.50 120 1M Open weights
23 Qwen3.8 27B Open weightsAlibaba Qwen — 33.7 — — 75.9 $0.10 $0.40 180 262K Open weights
24 MiniMax M3 Open weightsMiniMax — 29.2 90.9 1.0 — $0.30 $1.20 126 1M Open weights
25 Gemma 4 31B Open weightsGoogle — 19.0 75.8 — 58.4 $0.10 $0.40 200 262K Open weights
26 GPT-6 LunaOpenAI — — — — 82.5 $0.10 $0.50 130 1.05M Closed
27 Gemini 3.8 FlashGoogle — — 95.4 — 72.7 $0.75 $3.75 160 1.05M Closed
28 gpt-oss-120b Open weightsOpenAI — — 75.8 — — $0.05 $0.20 220 131K Open weights
29 Step 5 PreviewStepFun — — — — 76.9 $0.50 $2.00 95 1M Closed
30 Nemotron 3 Ultra Open weightsNVIDIA — — — — 62.8 $0.40 $1.60 70 1M Open weights

Why there are two intelligence columns

The AA Intelligence Index moved from v4.3.2 to v5 on 2026-10-27 and the scoring changed. The same DeepSeek V4.1 Flash scores 45.1 on the old version and 33.0 on the new one. The two sets must be read separately; mixing them produces completely wrong conclusions.

A dash is not a zero

It means the benchmark has no published result for that model, not that it scored zero. Open models and freshly released models are more likely to be missing data, so watch for sampling bias when comparing composites.

Prices are list price

All figures are first-party API list prices in USD per million tokens, excluding volume discounts and enterprise agreements. Some models are on promo pricing — see the pricing page for expiry dates.