$

llm-rate

Tuesday, 23 June 2026

In the last 24 hours we dispatched 1,556 tasks across 4 models. Here's what we picked, and why.

What we ran

An autonomous AI fleet, written in TypeScript, picks a model per task using a complexity router. No vibes, no PR team. This is the actual production output of that router:

ModelDispatchesShareWhy this one
01 claude-sonnet-4-6 1,066 68.5% implementation (standard)
02 claude-haiku-4-5 278 17.9% implementation (light)
03 gpt-5.4-mini 189 12.1% implementation (codex pool)
04 claude-opus-4-6 23 1.5% implementation (high complexity)

Window: 24h to 2026-05-16T00:00:00Z. Source: daemon routing logs. The router writes a decision per dispatch; we parsed 1556 of them.

If you don't have a router, here are the picks per common task

Filtered from arena.ai's leaderboard plus published API prices. Filter thresholds are listed under each tab; arguable. Treat this as a starting shortlist, not a verdict.

High volume, schemas, low complexity. Reliability + cheap.

Best value

qwen3-235b-a22b-thinking-2507

Alibaba · quality 1413.6 · $0.10/M blended

Best quality

mimo-v2.5-pro

Xiaomi · quality 1461.9 · $0.74/M blended

Filter: Blended price ≤ $2/M, quality ≥ 1250. Sorted by value. These workloads run at scale; small price diffs are real money. 106 models survived.

Model Quality Ctx In /1M Out /1M Value ↓
01 qwen3-235b-a22b-thinking-2507valueAlibaba 1413.6 262k $0.10 $0.10 413600.0
02 granite-4.1-8bIBM 1292.5 131k $0.05 $0.10 344070.6
03 gemma-3-4b-itGoogle 1290.8 131k $0.05 $0.10 342129.4
04 gemma-3n-e4b-itGoogle 1306.2 33k $0.06 $0.12 300235.3
05 deepseek-v4-flashDeepSeek 1429.6 1.0M $0.09 $0.18 280790.8
06 gemma-3-12b-itGoogle 1334.2 131k $0.05 $0.15 278516.7
07 gpt-oss-20bOpenAI 1287.7 131k $0.03 $0.14 269672.0
08 gpt-oss-120bOpenAI 1365.5 131k $0.04 $0.18 265410.3
09 gemma-3-27b-itGoogle 1358.2 131k $0.08 $0.16 263389.7
10 qwen3-30b-a3b-instruct-2507Alibaba 1383.8 131k $0.05 $0.19 256618.5
11 nvidia-nemotron-3-nano-30b-a3b-bf16Nvidia 1349.1 262k $0.06 $0.24 187682.8
12 mimo-v2.5Xiaomi 1426.1 1.0M $0.14 $0.28 179025.2
13 mimo-v2.5-proqualityXiaomi 1461.9 1.0M $0.43 $0.87 62461.1
14 deepseek-v4-proDeepSeek 1448.3 1.0M $0.43 $0.87 60619.3
15 deepseek-v4-pro-thinkingDeepSeek 1447.2 1.0M $0.43 $0.87 60478.7

What this is, and isn't

Right now this is filter-on-arena.ai plus a public log of what we ran. Arena Elo measures pairwise human preference on short prompts. It does not measure: whether a model produces valid JSON under a schema, whether it hallucinates function names, whether it refuses queries it shouldn't, latency p99, rate-limit behaviour. Production teams need those signals.

We're building a benchmark runner — fixed prompt suites for RAG, structured extraction, code refactoring, function calling — run daily against every model. Raw inputs, outputs, judge rationale, costs published. When that lands, the "picks" section gets its real backing. Until then, the picks section is opinion with a citation, not measurement.