Model Ratings.

The tracked cohort, scored and tiered.

62 models ranked against the runs local recipe · every axis sourced.

Scores are relative to the tracked cohort — swap the recipe or the field and the ranks move. The weights are an editorial call; the axis values underneath are sourced, cell by cell.

Snapshot · verified 1w ago · 2026-10-05

Recipe

Open-weight, fits your GPU — VRAM and context weighted alongside smarts.

The board · ranked · Runs local

Each cell brightens toward the cohort leader and is ringed when it holds it; dimmed rows are sparse (low coverage) — read their composite with care.

1
Claude Opus 5.5
Anthropic
40% · mediumS98
1. VRAM—
2. AA Index58
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: context
2
Claude Sonnet 5.5
Anthropic
40% · mediumS95
1. VRAM—
2. AA Index56
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
3
Claude Opus 5
Anthropic
60% · mediumS91
1. VRAM—
2. AA Index51
3. SWE-bench96%
4. LiveCode—
5. Context1M
Strongest: swebench · Weakest: aa_index
4
Claude Fable 5.1
Anthropic
40% · mediumS89
1. VRAM—
2. AA Index53
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
5
Gemini 4 Argon
Google DeepMind
40% · mediumS89
1. VRAM—
2. AA Index53
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
6
GPT-6 Astra
OpenAI
40% · mediumS89
1. VRAM—
2. AA Index53
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
7
Claude Fable 5
Anthropic
60% · mediumS88
1. VRAM—
2. AA Index50
3. SWE-bench95%
4. LiveCode—
5. Context1M
Strongest: swebench · Weakest: aa_index
8
GPT-6.1 Sol
OpenAI
40% · mediumS87
1. VRAM—
2. AA Index52
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
9
GPT-6 Sol
OpenAI
40% · mediumA81
1. VRAM—
2. AA Index48
3. SWE-bench—
4. LiveCode—
5. Context1.1M
Strongest: aa_index · Weakest: aa_index
10
Muse Spark 1.3
Meta
40% · mediumA80
1. VRAM—
2. AA Index48
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
11
GPT-5.6 Sol
OpenAI
40% · mediumA78
1. VRAM—
2. AA Index47
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
12
GLM-5.3-Flash
Z.AI (Zhipu)
60% · mediumA78
1. VRAM177 GB
2. AA Index42
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
13
gpt-oss-20b
OpenAI
40% · mediumA74
1. VRAM13 GB
2. AA Index—
3. SWE-bench—
4. LiveCode—
5. Context131K
Strongest: vram_q4 · Weakest: context
14
Mistral Small 3.2 24B
Mistral AI
40% · mediumA74
1. VRAM14 GB
2. AA Index—
3. SWE-bench—
4. LiveCode—
5. Context131K
Strongest: vram_q4 · Weakest: context
15
Qwen3.8 Max (0902)
Alibaba (Qwen)
40% · mediumA74
1. VRAM—
2. AA Index45
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
16
Gemma 3 27B
Google
40% · mediumA74
1. VRAM16 GB
2. AA Index—
3. SWE-bench—
4. LiveCode—
5. Context131K
Strongest: vram_q4 · Weakest: context
17
MiMo-V2.6-Flash
Xiaomi
60% · mediumA74
1. VRAM171 GB
2. AA Index38
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
18
GLM-5.3
Z.AI (Zhipu)
60% · mediumA74
1. VRAM414 GB
2. AA Index45
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
19
Step 5 Preview
StepFun
40% · mediumA72
1. VRAM—
2. AA Index44
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
20
gpt-oss-120b
OpenAI
40% · mediumA72
1. VRAM63 GB
2. AA Index—
3. SWE-bench—
4. LiveCode—
5. Context131K
Strongest: vram_q4 · Weakest: context
21
Qwen3-32B
Alibaba (Qwen)
40% · mediumA72
1. VRAM20 GB
2. AA Index—
3. SWE-bench—
4. LiveCode—
5. Context41K
Strongest: vram_q4 · Weakest: context
22
Phi-4
Microsoft
40% · mediumA71
1. VRAM9 GB
2. AA Index—
3. SWE-bench—
4. LiveCode—
5. Context16K
Strongest: vram_q4 · Weakest: context
23
DeepSeek V4 Flash Vision Exp
DeepSeek
60% · mediumA71
1. VRAM167 GB
2. AA Index35
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
24
MiMo-V2.6-Pro
Xiaomi
60% · mediumA70
1. VRAM563 GB
2. AA Index46
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: vram_q4
25
DeepSeek V4 Flash 0731
DeepSeek
60% · mediumA70
1. VRAM167 GB
2. AA Index34
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
26
Qwen3.8-Flash-Next
Alibaba (Qwen)
80% · highB70
1. VRAM99 GB
2. AA Index40
3. SWE-bench—
4. LiveCode91.9%
5. Context262K
Strongest: vram_q4 · Weakest: context
27
GPT-5.6 Terra
OpenAI
40% · mediumB69
1. VRAM—
2. AA Index42
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
28
Claude Opus 4.8
Anthropic
60% · mediumB69
1. VRAM—
2. AA Index42
3. SWE-bench88.6%
4. LiveCode—
5. Context1M
Strongest: swebench · Weakest: aa_index
29
Gemini 3.8 Flash
Google DeepMind
40% · mediumB67
1. VRAM—
2. AA Index41
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: aa_index
30
Claude Opus 4.7
Anthropic
60% · mediumB66
1. VRAM—
2. AA Index41
3. SWE-bench87.6%
4. LiveCode—
5. Context1M
Strongest: swebench · Weakest: aa_index
31
DeepSeek V4.1 Flash
DeepSeek
60% · mediumB65
1. VRAM510 GB
2. AA Index39
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
32
Qwen3.8-27B
Alibaba (Qwen)
80% · highB65
1. VRAM15 GB
2. AA Index34
3. SWE-bench—
4. LiveCode90.3%
5. Context262K
Strongest: vram_q4 · Weakest: context
33
Gemini 3.7 Flash
Google DeepMind
40% · mediumB63
1. VRAM—
2. AA Index39
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
34
GPT-5.5
OpenAI
60% · mediumB62
1. VRAM—
2. AA Index38
3. SWE-bench88.7%
4. LiveCode—
5. Context922K
Strongest: swebench · Weakest: aa_index
35
GPT-6 Luna
OpenAI
40% · mediumB61
1. VRAM—
2. AA Index38
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
36
Grok 4.7
xAI
40% · mediumB60
1. VRAM—
2. AA Index46
3. SWE-bench—
4. LiveCode—
5. Context500K
Strongest: aa_index · Weakest: context
37
GPT-5.6 Luna
OpenAI
40% · mediumB59
1. VRAM—
2. AA Index37
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
38
DeepSeek V4 Flash (Preview)
DeepSeek
100% · highB59
1. VRAM95 GB
2. AA Index24
3. SWE-bench79%
4. LiveCode91.6%
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
39
DeepSeek V4 Pro
DeepSeek
100% · highB58
1. VRAM517 GB
2. AA Index30
3. SWE-bench80.6%
4. LiveCode93.5%
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
40
Grok 4.6
xAI
40% · mediumB56
1. VRAM—
2. AA Index44
3. SWE-bench—
4. LiveCode—
5. Context500K
Strongest: aa_index · Weakest: context
41
Gemini 3.6 Flash
Google DeepMind
40% · mediumC54
1. VRAM—
2. AA Index34
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
42
GLM-5.2
Z.AI (Zhipu)
40% · mediumC54
1. VRAM—
2. AA Index34
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
43
Gemini 3.5 Flash
Google DeepMind
40% · mediumC52
1. VRAM—
2. AA Index33
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
44
DeepSeek V4 Pro 0813
DeepSeek
60% · mediumC50
1. VRAM908 GB
2. AA Index36
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: vram_q4 · Weakest: aa_index
45
Grok 4.5
xAI
40% · mediumC47
1. VRAM—
2. AA Index39
3. SWE-bench—
4. LiveCode—
5. Context500K
Strongest: aa_index · Weakest: context
46
MiniMax-M3
MiniMax
40% · mediumC45
1. VRAM—
2. AA Index29
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
47
Gemini 3.1 Pro
Google DeepMind
60% · mediumC42
1. VRAM—
2. AA Index30
3. SWE-bench80.6%
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
48
Claude Sonnet 4.6
Anthropic
60% · mediumC41
1. VRAM—
2. AA Index30
3. SWE-bench80.2%
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
49
Qwen3.8 Max
Alibaba (Qwen)
60% · mediumC41
1. VRAM1345 GB
2. AA Index40
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: vram_q4
50
Kimi K3
Moonshot AI
60% · mediumD40
1. VRAM1529 GB
2. AA Index44
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: aa_index · Weakest: vram_q4
51
Claude Sonnet 5
Anthropic
60% · mediumD39
1. VRAM—
2. AA Index38
3. SWE-bench72.7%
4. LiveCode—
5. Context1M
Strongest: context · Weakest: swebench
52
MiMo-V2.5-Pro
Xiaomi
40% · mediumD39
1. VRAM—
2. AA Index26
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
53
Grok 4.3
xAI
40% · mediumD37
1. VRAM—
2. AA Index25
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
54
Qwen3.7 Plus
Alibaba (Qwen)
40% · mediumD37
1. VRAM—
2. AA Index25
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
55
Kimi K2.6
Moonshot AI
80% · highD33
1. VRAM—
2. AA Index27
3. SWE-bench80.2%
4. LiveCode89.6%
5. Context256K
Strongest: livecodebench · Weakest: aa_index
56
GPT-5.3 Codex
OpenAI
40% · mediumD33
1. VRAM—
2. AA Index33
3. SWE-bench—
4. LiveCode—
5. Context400K
Strongest: aa_index · Weakest: aa_index
57
Qwen3.7 Max
Alibaba (Qwen)
60% · mediumD32
1. VRAM—
2. AA Index29
3. SWE-bench—
4. LiveCode78.2%
5. Context1M
Strongest: context · Weakest: livecodebench
58
Gemini 3.5 Flash-Lite
Google DeepMind
40% · mediumD32
1. VRAM—
2. AA Index22
3. SWE-bench—
4. LiveCode—
5. Context1M
Strongest: context · Weakest: aa_index
59
Muse Spark
Meta
40% · mediumD25
1. VRAM—
2. AA Index31
3. SWE-bench—
4. LiveCode—
5. Context262K
Strongest: aa_index · Weakest: context
60
GPT-5.4 mini
OpenAI
40% · mediumD16
1. VRAM—
2. AA Index24
3. SWE-bench—
4. LiveCode—
5. Context400K
Strongest: context · Weakest: aa_index
61
Kimi K2.7 Code
Moonshot AI
40% · mediumD15
1. VRAM—
2. AA Index26
3. SWE-bench—
4. LiveCode—
5. Context256K
Strongest: context · Weakest: aa_index
62
GLM-5.1
Z.AI (Zhipu)
40% · mediumD13
1. VRAM—
2. AA Index26
3. SWE-bench—
4. LiveCode—
5. Context200K
Strongest: aa_index · Weakest: aa_index