SShogo Bench

Agent Index v1.0

The Shogo Agent Index

A weighted index of 17 benchmarks that measure what agents actually do. 92 models are ranked: 44 have enough measured results to be Supported; the rest are Estimated from at least two measured agentic results plus overall capability. 147 more models have a capability score but too few agentic results to rank yet. Full methodology.

92 models
ModelAgent IndexCodingLong-hznKnow. workComp. useSec/ctxMeasuredAgent $/MContext
Claude Fable 5.2
AnthropicSupported
76.4
71.7–81.2
867867776532%$3.211M
Gemini 4 Carbon
GoogleSupported
73.6
67.2–80
828063686327%$0.841M
Claude Opus 5.5
AnthropicSupported
70.9
64.8–76.9
817658666232%$1.691M
GPT-6 Astra
OpenAISupported
70.9
66.5–75.4
808056676338%$4.621.1M
GPT-6.1 Sol
OpenAIEstimated
70.0
62.2–77.7
807854666218%$0.841.1M
Gemini 4 Argon
GoogleSupported
69.5
62.9–76.2
777859616132%$0.841M
Gemini 4 Hydrogen
GoogleEstimated
69.4
61.8–77.1
787858616123%$0.631M
GPT-6.1 Terra
OpenAIEstimated
69.0
60.5–77.5
797456636214%$0.421.1M
Claude Sonnet 5.5
AnthropicEstimated
69.0
61.3–76.7
797557616123%$0.841M
Claude Fable 5.1
AnthropicSupported
68.2
62.7–73.7
807155626136%$4.011M
Claude Opus 5
AnthropicSupported
65.4
59.8–71
747452565939%$2.311M
Claude Fable 5
AnthropicSupported
64.1
59–69.3
746952555938%$4.621M
GPT-6 Sol
OpenAIEstimated
63.6
54.2–73
727648545914%$0.921.1M
GPT-5.6 Sol
OpenAISupported
61.8
57.1–66.5
717246535843%$1.851.1M
Muse Wonder
MetaEstimated
60.9
51.9–69.9
687050505722%$1.691M
GPT-5.6 Terra
OpenAIEstimated
58.8
50.7–66.9
666947485624%$1.021.1M
DeepSeek V4.1 Flash 1017
DeepSeek· openEstimated
57.6
44.4–70.8
636649455412%$0.0501M
Claude Haiku 5.5
AnthropicEstimated
57.5
48.4–66.5
676246475523%$0.0501M
GPT-5.5
OpenAISupported
56.7
52.5–61
636746435560%$2.551.1M
Claude Opus 4.8
AnthropicSupported
56.3
50.9–61.8
636643475544%$2.311M
Gemini 3.8 Flash
GoogleSupported
56.2
48.7–63.7
636447445334%$0.351M
Grok 4.6
xAISupported
56.2
48.3–64
616847425229%$0.98500K
Gemini 3.7 Flash
GoogleEstimated
55.7
44.2–67.1
616448435324%$0.351M
Inkling 2 Small
thinkingmachines· openEstimated
55.2
42.3–68.2
626642445413%$0.141M
GLM-5.3
Z.ai (Zhipu)· openSupported
55.2
47.7–62.7
626843405133%$0.611M
Kimi K3
Moonshot AI· openSupported
54.9
47.3–62.5
636542435329%$1.391M
Grok 4.7
xAIEstimated
54.8
45.1–64.5
586947395023%$0.98500K
Muse Spark 1.3
MetaEstimated
54.7
41.6–67.7
606445435317%$0.0301M
Claude Opus 4.7
AnthropicSupported
54.6
47.7–61.5
606941445244%$2.311M
Qwen3.8 Max Preview
Alibaba (Qwen)Estimated
53.6
41.7–65.5
586346415219%Free1M
Claude Opus 4.6
AnthropicSupported
53.2
47.9–58.6
576741395764%$2.311M
Claude Sonnet 5
AnthropicSupported
52.4
45.1–59.8
576543405128%$0.921M
GPT-5.4
OpenAISupported
51.6
47.2–56.1
576140425464%$1.271.1M
GPT-5.6 Luna
OpenAIEstimated
51.4
42.8–59.9
596139405124%$0.101.1M
DeepSeek V4.1 Flash
DeepSeek· openEstimated
51.2
37.1–65.2
576239405113%$0.121M
Grok 4.5
xAIEstimated
49.4
40.9–57.9
545942364824%$0.81500K
GPT-5.3 Codex
OpenAISupported
48.4
38.5–58.4
536237354732%$1.06400K
GLM-5.3-Flash
Z.ai (Zhipu)· openSupported
48.3
36.5–60
545740344727%$0.161M
Gemini 3.6 Flash
GoogleEstimated
48.3
35.5–61.2
535839374914%$0.351M
Muse Spark 1.1
MetaEstimated
48.3
39.3–57.2
536434364820%$0.501M
Muse Spark 1.2
MetaEstimated
48.2
35.8–60.6
535838374919%$0.501M
GLM-5.2
Z.ai (Zhipu)· openSupported
47.3
39.3–55.3
496637334532%$0.611M
Gemini 3.5 Flash
GoogleSupported
47.3
39–55.6
526233374928%$0.76200K
GPT-5.2
OpenAISupported
46.4
39.5–53.2
495937364658%$0.96400K
Gemini 3.1 Pro Preview
GoogleSupported
46.0
41.6–50.5
475936394964%$1.021M
Claude Opus 4.5 (latest)
AnthropicSupported
46.0
41.5–50.6
476137305267%$2.31200K
Claude Sonnet 4.6
AnthropicSupported
44.9
39.6–50.3
436338344551%$1.391M
Gemini 3 Pro Preview
GoogleSupported
44.7
37.9–51.6
486034334358%$2.481M
Grok 4.20 (Non-Reasoning)
xAIEstimated
43.9
33.3–54.5
455934324519%$0.462M
Kimi K2.6
Moonshot AI· openEstimated
43.7
32.3–55.1
466133294320%$0.46262K
Gemini 3 Flash Preview
GoogleSupported
43.7
34.9–52.5
465735324431%$0.211M
GLM-5.1
Z.ai (Zhipu)· openEstimated
42.5
32.6–52.4
435934294222%$0.46200K
Kimi K2.7 Code
Moonshot AI· openEstimated
42.1
33.1–51.2
435933304220%$0.96262K
DeepSeek V4 Pro
DeepSeek· openEstimated
40.7
29.3–52.1
435532284114%$0.501M
Qwen3.6 Max Preview
Alibaba (Qwen)Estimated
40.5
29.2–51.9
425631274014%$1.08262K
GPT-5.1 Codex Max
OpenAIEstimated
39.6
26.4–52.9
434832284118%$0.67400K
GPT-5.1
OpenAISupported
38.8
30.4–47.3
404932294436%$0.69400K
Claude Sonnet 4.5 (latest)
AnthropicSupported
38.7
34.6–42.8
385432273962%$1.39200K
Qwen3.6 Plus
Alibaba (Qwen)Estimated
38.0
27.1–48.9
365730263919%Free1M
GPT-5
OpenAISupported
37.8
27.8–47.8
404731274047%$0.76400K
MiniMax-M3
MiniMax· openEstimated
36.7
26.5–46.8
374931243720%$0.15524K
GLM-5
Z.ai (Zhipu)· openSupported
36.5
27.7–45.4
385427213527%$0.33205K
Kimi K2.5
Moonshot AI· openSupported
36.5
28.3–44.8
384530283931%$0.31262K
Inkling
thinkingmachines· openEstimated
35.8
23–48.6
384430253813%$0.79262K
Qwen3.5 Plus
Alibaba (Qwen)Estimated
32.7
21.5–44
363728233711%$0.201M
Claude Opus 4.1 (latest)
AnthropicSupported
32.1
25.4–38.8
344030192943%$5.96197K
o3
OpenAISupported
31.4
22.5–40.2
344227113543%$0.97200K
Grok 4
xAISupported
30.9
22.4–39.4
324225223035%——
GPT-5 Mini
OpenAIEstimated
30.4
21.4–39.3
323626213322%$0.15264K
Kimi K2 Thinking
Moonshot AI· openEstimated
30.3
19.8–40.9
323726203323%$0.35262K
GLM-4.7
Z.ai (Zhipu)· openEstimated
30.0
21.3–38.8
304723162919%$0.24205K
DeepSeek V3.2
DeepSeek· openSupported
30.0
22.1–37.8
324124182927%$0.16128K
o4-mini
OpenAIEstimated
29.8
19.2–40.4
313924193221%$0.53200K
MiniMax-M2.5
MiniMax· openEstimated
28.4
19.5–37.2
323424172719%Free197K
Claude-Opus-4
AnthropicSupported
27.9
20.4–35.4
293824162732%$5.96193K
Claude Haiku 4.5
AnthropicEstimated
27.1
18.9–35.3
303423152819%$0.46200K
Qwen3 Coder 480B A35B
Alibaba (Qwen)· openEstimated
26.4
16.9–35.9
273622152712%$0.17262K
Claude Sonnet 4
AnthropicSupported
26.2
19.5–33
263623192528%$1.20983K
Qwen3 Max
Alibaba (Qwen)Estimated
25.7
16.2–35.3
273323152711%$0.18262K
Kimi K2 0905
Moonshot AI· openEstimated
25.3
16.2–34.3
263421132512%$1.16262K
Gemini 2.5 Pro
GoogleSupported
24.5
19.4–29.6
273221132453%$0.761M
Claude 3.7 Sonnet
AnthropicSupported
24.2
18.6–29.8
253521151836%$1.20197K
MiniMax-M2
MiniMax· openEstimated
23.9
16–31.9
263120122314%$0.40400K
Gemini 2.5 Flash
GoogleEstimated
23.6
15.5–31.8
243321132414%$0.191M
o3-mini
OpenAIEstimated
23.3
14.9–31.6
24342113199%$1.13200K
GPT OSS 120B
OpenAI· openSupported
22.1
15.7–28.5
233117122432%$0.040131K
DeepSeek-R1
DeepSeek· openEstimated
21.1
13.5–28.6
213118102115%$0.7180K
o1
OpenAIEstimated
16.7
11.7–21.7
16281661115%$15.90200K
Claude-Sonnet-3.5
AnthropicEstimated
14.8
10.9–18.8
14251441419%$1.20189K
GPT-4o (2024-08-06)
OpenAISupported
12.2
10–14.5
10221221134%$1.68128K
Qwen2.5-72B
Alibaba (Qwen)· openEstimated
12.1
8.5–15.7
1120122914%——
Claude 3 Opus
AnthropicEstimated
11.0
7.9–14.1
11161221015%——

Tick models to compare them side by side. Pillar scores in white include measured results; grey ones are fully estimated. Shaded band = 95% range.

What’s in the index

Agentic coding

35%

Fixing real repos, working in a terminal and shipping multi-file changes without a human in the loop.

Long-horizon autonomy

20%

Staying on task for hours: long software tasks, running a business, training a model end to end.

Knowledge-work agents

25%

Professional deliverables graded by experts: banking, consulting, law, research reports, paid freelance jobs.

Computer use

10%

Operating a real desktop and browser through screenshots, mouse and keyboard.

Security & context learning

10%

Capture-the-flag security tasks and learning new rules from long context on the fly.