Alternatives to Inkling 2 Small

Thinking Machines LMArena 1402.6 $0.475/M on a balanced workload prices as of 2026-11-22

Every higher-scoring or cheaper alternative has a recorded capability loss.

Of the 143 models carrying both an independent score and a published price at the same delivery mode, 17 score at least as high as Inkling 2 Small and cost no more under at least one listed workload, with at least one of those dimensions improved. 17 have a recorded capability loss, named on every row.

Dominance is not a property of a model. It is a property of a model, a capability lens and a workload, and all three are named on every row. Scores are LMArena Elo, used under CC BY 4.0 from the official dataset; prices are blended per million tokens. A higher score is not a drop-in replacement.

This page lists models only while a candidate improves score or price without worsening the other under a listed workload. When that comparison no longer holds, this address stays and says so.

Where it holds

WorkloadInkling 2 Small $/MRecorded capabilities preservedRecorded capability losses
Balanced$0.475015
Summarise$0.258013
Chat$0.499013
Code gen$0.698013
Agentic$0.244011

Workload mixes are defined on the methodology page. Cached input is priced at the cached rate, and reasoning tokens at the worse of the reasoning and output rates.

What you would be replacing

Intelligence1402.6
Context window1M
Max output943K
Input modestext, image
Tool useyes
Extended reasoningyes

Preserving recorded capabilities requires matching or beating these fields. Full record for Inkling 2 Small, or check Inkling 2 Small against the whole catalogue.

Alternatives with recorded capability losses 17

These improve score or price without worsening the other under the listed workloads, but reduce at least one recorded capability. Read the last column before switching.

ModelIntelligence$/MHolds underRecorded capability loss
GLM 5.3 Flash Z.ai1471.4 +68.8$0.148
  • Balanced −69%
  • Summarise −63%
  • Chat −68%
  • Code gen −70%
  • Agentic −59%
  • 944K → 131K max output
DeepSeek V4.1 Flash DeepSeek1463 +60.4$0.227
  • Balanced −52%
  • Summarise −61%
  • Chat −45%
  • Code gen −44%
  • Agentic −50%
  • 944K → 393K max output
MiMo-V2.6-Flash Xiaomi1449.8 +47.2$0.175
  • Balanced −63%
  • Summarise −58%
  • Chat −69%
  • Code gen −69%
  • Agentic −68%
  • 944K → 131K max output
Gemma 4 31B Google1442.5 +39.9$0.205
  • Balanced −57%
  • Summarise −41%
  • Chat −51%
  • Code gen −58%
  • Agentic −27%
  • 1M → 262K context
  • 944K → 16K max output
Hy3 Tencent1439 +36.4$0.231
  • Balanced −51%
  • Summarise −52%
  • Chat −48%
  • Code gen −48%
  • Agentic −46%
  • 1M → 262K context
  • 944K → 236K max output
  • no image input
Gemma 4 26B A4B Google1434.2 +31.6$0.123
  • Balanced −74%
  • Summarise −73%
  • Chat −70%
  • Code gen −72%
  • Agentic −66%
  • 1M → 262K context
  • 944K → 236K max output
DeepSeek V4 Flash 0423 DeepSeek1431.9 +29.3$0.113
  • Balanced −76%
  • Summarise −71%
  • Chat −79%
  • Code gen −80%
  • Agentic −75%
  • 944K → 393K max output
  • no image input
GPT-5.6 Luna OpenAI1430.9 +28.3$0.450
  • Balanced −5%
  • Summarise −23%
  • Agentic −1%
  • 944K → 128K max output
DeepSeek V3.2 DeepSeek1424.6 +22$0.394
  • Balanced −17%
  • Summarise −2%
  • Chat −12%
  • Code gen −18%
  • 1M → 164K context
  • 944K → 147K max output
  • no image input
DeepSeek V3.2 Exp DeepSeek1422.3 +19.7$0.305
  • Balanced −36%
  • Chat −35%
  • Code gen −49%
  • 1M → 164K context
  • 944K → 147K max output
  • no image input
Qwen3 235B A22B Instruct 2507 Qwen1419.2 +16.6$0.205
  • Balanced −57%
  • Summarise −56%
  • Chat −45%
  • Code gen −48%
  • Agentic −35%
  • 1M → 262K context
  • 944K → 118K max output
  • no image input
  • no extended reasoning
Qwen3 VL 235B A22B Instruct Qwen1420 +17.4$0.370
  • Balanced −22%
  • Summarise −19%
  • Chat −11%
  • Code gen −14%
  • 1M → 262K context
  • 944K → 33K max output
  • no extended reasoning
DeepSeek V3.1 DeepSeek1419.8 +17.2$0.453
  • Balanced −5%
  • 1M → 164K context
  • 944K → 147K max output
  • no image input
Qwen3 Next 80B A3B Instruct Qwen1417.2 +14.6$0.343
  • Balanced −28%
  • Summarise −46%
  • Chat −1%
  • Code gen under 1% less
  • Agentic −1%
  • 1M → 262K context
  • 944K → 236K max output
  • no image input
  • no extended reasoning
DeepSeek V3.1 Terminus DeepSeek1417.3 +14.7$0.453
  • Balanced −5%
  • 1M → 164K context
  • 944K → 147K max output
  • no image input
MiMo-V2.6-Pro Xiaomi1492.7 +90.1$0.478 chat
  • Chat −4%
  • Code gen −6%
  • Agentic −1%
  • 944K → 131K max output
Gemini 3.1 Flash Lite Preview Google1416.2 +13.6$0.248 summarise
  • Summarise −4%
  • 944K → 66K max output

What this compares, and what it leaves out

  • Quality is LMArena Elo, used under CC BY 4.0 from the official dataset. The 95% confidence interval on a difference between two scores is about ±10.46 points, so a gap smaller than that is marked tie rather than an improvement — see significance bands.
  • 199 further models at this delivery mode carry a price but no independent score. They are absent from the comparison in both directions — unrated is not a zero, and an unmeasured model is neither an alternative nor a worse buy.
  • Retired models are never offered as an alternative, and a model only competes against its own delivery mode: batch trades latency for price, so it is not a like-for-like swap.
  • Nothing here measures latency, throughput, rate limits or how a model behaves on your prompts. Two models with the same index score are not interchangeable.
  • Models with no higher-scoring or cheaper alternative that is no worse on the other axis are on the value frontier. Models with such an alternative are listed here.
Evidence & Ask