ModelSpec is an analysis of alternatives for AI models: it decides which models fit a job from your requirements, sourced benchmarks and real cost, and shows its work.

General assistant, balanced

Find an active assistant that balances how people rate its responses against task cost.

Worked example as of 2026-09-23, from the public data image.

Response quality first, cost as a real second term.

Requirements

Preferences

Applied template spec
spec_version
1
where
  • model.class = text-generator
  • model.lifecycle = active
optimize
weights
chat_preference
0.6
-offering.cost_per_task
0.4

Worked answer

Decision status: partial.

Tied on the evidence

probability bands: a model joins the leader when P(its score >= the leader's) >= 0.25 under the score posteriors; a weighted capability with an 80% interval wider than 2.8 is not enough evidence

Tie-breakers describe members of the tie. They do not establish a winner.

This is a partial answer. Unknown facts or missing objective evidence can leave eligible candidates unranked.

probability bands: a model joins the leader when P(its score >= the leader's) >= 0.25 under the score posteriors; a weighted capability with an 80% interval wider than 2.8 is not enough evidence

Score and score interval use the objective's dimensionless normalised units. Cost per task is for the engine-selected offering and the applied spec's task size. When task size is omitted, the engine uses its documented default. P(best) is the engine probability. Unavailable means the engine did not return a value. Thin evidence is undetermined and is not a recommendation.

Best band

ModelScoreScore intervalP(best)Cost per task, USDCapability estimates and intervals
anthropic/claude-opus-5-50.846914[0.703041, 0.990787]0.41406250.24
  • chat_preference: 1.584232494; interval [0.565758585, 2.602706403]; benchmarks 1; direct benchmarks 1
openai/gpt-6-sol0.790483[0.64661, 0.934356]0.207031250.12
  • chat_preference: 0.602126414; interval [-0.416347496, 1.620600323]; benchmarks 1; direct benchmarks 1
qwen/qwen3-8-max-09020.799797[0.655924, 0.94367]0.191406250.104
  • chat_preference: 0.59037787; interval [-0.428096039, 1.60885178]; benchmarks 1; direct benchmarks 1
xai/grok-4-70.745338[0.601465, 0.889211]0.074218750.104
  • chat_preference: 0.204866691; interval [-0.813607219, 1.2233406]; benchmarks 1; direct benchmarks 1
google/gemini-3-8-flash0.822556[0.7976, 0.847512]0.07031250.045
  • chat_preference: 0.465029132; interval [0.288363151, 0.641695113]; benchmarks 24; direct benchmarks 24
google/gemini-3-7-flash0.813658[0.788701, 0.838614]0.042968750.045
  • chat_preference: 0.402038706; interval [0.225372725, 0.578704687]; benchmarks 24; direct benchmarks 24

Rest band

ModelScoreScore intervalP(best)Cost per task, USDCapability estimates and intervals
meta/muse-spark-1-30.765089[0.7325, 0.797678]0.00.067
  • chat_preference: 0.165034925; interval [-0.065661987, 0.395731837]; benchmarks 13; direct benchmarks 13
zhipu/glm-5-30.758923[0.736209, 0.781637]0.00.0736
  • chat_preference: 0.153431285; interval [-0.007361761, 0.31422433]; benchmarks 29; direct benchmarks 29
google/gemini-3-1-pro-preview0.701498[0.680513, 0.722483]0.00.128
  • chat_preference: 0.01105026; interval [-0.137500042, 0.159600562]; benchmarks 34; direct benchmarks 34
meta/muse-spark-1-10.699813[0.671785, 0.727841]0.00.067
  • chat_preference: -0.297052895; interval [-0.4954636, -0.098642189]; benchmarks 19; direct benchmarks 19
anthropic/claude-opus-4-60.691491[0.671096, 0.711887]0.00.3
  • chat_preference: 0.775317472; interval [0.630936802, 0.919698141]; benchmarks 36; direct benchmarks 36
google/gemini-2-5-flash0.69117[0.668829, 0.713511]0.00.022
  • chat_preference: -0.576719849; interval [-0.734872721, -0.418566977]; benchmarks 34; direct benchmarks 34
google/gemini-3-5-flash0.683543[0.662558, 0.704528]0.00.096
  • chat_preference: -0.271422803; interval [-0.419973105, -0.122872501]; benchmarks 34; direct benchmarks 34
zhipu/glm-5-20.679011[0.656296, 0.701725]0.00.0736
  • chat_preference: -0.412265302; interval [-0.573058347, -0.251472257]; benchmarks 29; direct benchmarks 29
anthropic/claude-opus-4-70.676242[0.655258, 0.697227]0.00.3
  • chat_preference: 0.667370368; interval [0.518820066, 0.81592067]; benchmarks 34; direct benchmarks 34
anthropic/claude-opus-50.65577[0.634786, 0.676755]0.00.3
  • chat_preference: 0.522450073; interval [0.373899771, 0.671000375]; benchmarks 34; direct benchmarks 34
openai/gpt-5-40.655352[0.634956, 0.675747]0.00.16
  • chat_preference: -0.16025079; interval [-0.304631459, -0.01587012]; benchmarks 36; direct benchmarks 36
openai/gpt-5-6-sol0.654065[0.63276, 0.67537]0.00.24
  • chat_preference: 0.219061677; interval [0.068242727, 0.369880626]; benchmarks 33; direct benchmarks 33
deepseek/deepseek-v4-pro0.62854[0.605826, 0.651254]0.00.06864
  • chat_preference: -0.793630284; interval [-0.954423329, -0.632837239]; benchmarks 29; direct benchmarks 29
anthropic/claude-fable-50.522509[0.501524, 0.543494]0.00.6
  • chat_preference: 1.035674964; interval [0.887124662, 1.184225266]; benchmarks 34; direct benchmarks 34
anthropic/claude-fable-5-10.485654[0.463682, 0.507627]0.00.6
  • chat_preference: 0.774780684; interval [0.619237933, 0.930323435]; benchmarks 31; direct benchmarks 31
openai/gpt-6-astra0.381863[0.35788, 0.405845]0.00.6
  • chat_preference: 0.040043976; interval [-0.12972742, 0.209815373]; benchmarks 26; direct benchmarks 26

Unknown requirements

Some candidates may qualify. Missing facts:

Engine warnings

Omitted result rows: models 10; offerings 18. The bands above include all ranked models.

Decision dec_a6d258fd73f7d9848e50fa39 on public snapshot snap_05a963d2477ac2fb.

What this template teaches

A 60/40 blend lets a large price gap overturn a small quality lead.

The live board uses current data. Its answer can change as evidence changes.

Apply this template on Decide

Human lookup is free on Decide. CLI, API, and MCP access requires a key. Access and pricing.

Use-case examples · How we decide

Feedback