Compare / head-to-head

Inkling-SmallvsJamba2 Mini

Inkling-Small leads 1 of 1 shared benchmark. Jamba2 Mini is 2.5x cheaper per token. Inkling-Small has the larger context window (1M tokens).

Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.

Shared benchmarks
1 – 0
Inkling-Small leads
Cheaper per token
Jamba2 Mini
2.5x cheaper, input + output
Larger context
Inkling-Small
1M tokens
Providers
2 providers
Thinking Machines Lab · AI21 Labs
Inkling-Small
Thinking Machines Lab · released 2026-07-30
textvisionaudio
Context
1M
Max out
—
Input /1M
$0.3 ↗
Output /1M
$1.2 ↗
Cached /1M
$0.06
Scores
29 · 12 core
Jamba2 Mini
AI21 Labs · released 2026-01-08
text
Context
256K
Max out
4K
Input /1M
$0.2 ↗
Output /1M
$0.4 ↗
Cached /1M
—
Scores
4 · 0 core
Quality

Benchmark matrix

BenchmarkInkling-SmallJamba2 MiniΔEdge
Reported by both · 1
IFBench82.2% ↗38% ↗+44.2 ptInkling-Small
Only Inkling-Small reports · 27
SWE-bench Verified80.2% ↗not reported——
GPQA Diamond88.5% ↗epoch runnot reported——
AIME 202695.5% ↗not reported——
LMArena Elo1405.2 ↗not reported——
ARC-AGI-184% ↗not reported——
ARC-AGI-240.1% ↗not reported——
BrowseComp (with context management)77.4% ↗not reported——
CharXiv RQ77.4% ↗not reported——
CharXiv RQ (with python)81.3% ↗not reported——
Chess Puzzles (Epoch AI run)18% ↗epoch runnot reported——
FrontierMath Tier 4 v2 (Epoch AI run)17.1% ↗epoch runnot reported——
FrontierMath Tiers 1-3 v2 (Epoch AI run)46.3% ↗epoch runnot reported——
HMMT Feb 202690.2% ↗not reported——
LMArena Agent (LMArena)-0.1033 ↗not reported——
LMArena Vision (LMArena)1206.3 ↗not reported——
LMArena WebDev (LMArena)1408.2 ↗not reported——
MCP Atlas (all)79.2% ↗not reported——
MMAU77% ↗not reported——
Mystery Game Puzzles (Epoch AI run)6% ↗epoch runnot reported——
OTIS Mock AIME 2024-2025 (Epoch AI run)90% ↗epoch runnot reported——
SAGE (Vals AI)32.7% ↗not reported——
SimpleQA Verified20.6% ↗not reported——
SWE-bench Pro (public)55.9% ↗not reported——
Terminal-Bench 2.1 (best harness)64.7% ↗not reported——
Terminal-Bench 4.0 (Vals AI)0% ↗not reported——
Toolathlon Verified54.4% ↗not reported——
Toolathlon-Verified (HKUST)54.4% ↗not reported——
Only Jamba2 Mini reports · 3
Collienot reported33% ↗——
FACTSnot reported57% ↗——
IFEvalnot reported97% ↗——
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Inkling-Small minus Jamba2 Mini in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing

Side by side

SpecInkling-SmallJamba2 MiniEdge
Context window1M tokens256K tokensInkling-Small
Max output—4K tokens—
Input price / 1M$0.3 ↗$0.2 ↗Jamba2 Mini
Output price / 1M$1.2 ↗$0.4 ↗Jamba2 Mini
Cached input / 1M$0.06 ↗——
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$1.5$0.6Jamba2 Mini
Modalitiestext · vision · audiotextInkling-Small
Released2026-07-302026-01-08—
Cited benchmark scores294—
More matchups

Inkling-Small vs …

More matchups

Jamba2 Mini vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run Inkling-Small and Jamba2 Mini on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.